Faculty raises

At MSU, there is no annual inflation adjustment with salaries, and this results in a massive loss of revenue for faculty over the years. Here is a Shiny app that I created a couple of years back based on data that I FOIA’d from MSU.¹ The app shows the trends in MSU salaries over the period 2011-2023. (Note: the app runs on a free Shiny server that is bandwidth limited to a total of 25 hours per month across all users, so it might not work depending on usage.)

Instead of any inflation adjustment, MSU has “merit” raises. However, MSU doesn’t define what it means by “merit”; at least, it isn’t defined in the faculty handbook. Here is the extent of the discussion under “Annual Performance Review for Fixed Term Faculty and Academic Staff” from what I can find.

All units must have procedures for annual written evaluation of all fixed term faculty and academic staff to support the annual merit process and to provide a basis for a clear statement of performance expectations and accomplishments. It is recognized that provisions and practices in units may vary; however, all evaluation procedures must incorporate, at the minimum, the principles included in this policy and must be applied consistently to all fixed term faculty and academic staff within the unit.

Honestly speaking, given there is no definition of “merit”, I can’t understand why colleges/departments can’t institute merit increase rules that prioritise inflation adjustment and then give raises reflective of merit work above it. In fact, there is no verbiage that I could find that says explicitly that merit raises cannot have an inflation adjustment component. Perhaps, I haven’t looked hard enough. But, let’s put this aside for now.

Anyhow, I take it that the lack of a precise definition allows colleges/departments within MSU to define it as they see fit. Now, in my opinion, “merit”, to the extent that there is any fair value to the concept,² should be about whether one has done their contractually obligated work well. Instead, in many departments, there is a tendency to define it as how much work has one done? The former is a healthy interpretation of the contract that we have, while the latter sets up a rat race, where each faculty has to burn themselves out in research, additional teaching/training commitments outside class hours, crazy amounts of service, … I am all for improving quality of work, but it needs to respect contracts, otherwise it is abusive. And as far as I can see, currently the merit system at MSU tends to be abusive.

OK, let’s grant for now that that is how merit raises work, so let’s turn a blind eye to the abusive system in place. However, there is still a big issue that deeply bothers me. Nowhere in the faculty handbook is there a mention of how merit raises are calculated. Some departments use a percentage of the salary as a merit raise, while others use a more complicated flat raise+percentage raise approach. The question at hand in this blog post is what is an appropriate way to apportion raises (I say apportion as the college/department has a fixed amount of money to use for merit raises, given to them by MSU).

To be more specific, let’s say that two faculty have the same merit rating, one could ask, what should their raises be? There are three ways in which you could imagine this playing out: a regressive scheme, a flat scheme, and a progressive scheme.

First, a regressive salary raise scheme is one where higher earners get a larger amount. This is what a percentage-based merit raise system is, a regressive system. So, if we say everyone within a certain merit rating gets the same proportional increment, then of course, this benefits the high wage earners.

Second, a flat salary raise scheme is one where all earners within a merit rating receive the same amount of salary raise. This is, in a sense, a neutral system.

Third, a progressive salary raise scheme is one where the lower earners get a larger amount. This is the option that I strongly prefer. As enshrined in the famous Difference Principle³ by John Rawls, who wrote extensively about justice and fairness:

Social and economic inequalities are to be arranged so that they are to the greatest benefit of the least advantaged. [Rawls (1971/1999); p. 302 (original)/266 (revised)]

But, what I deem to be the fairest and most just system of raises is not even up for discussion anywhere at MSU, from what I know. Maybe, in a distant future, when fairness will really mean something substantive, it will make an appearance. For now, the discussion (if that) that follows is between flat raises and regressive raises (aka proportional raises). So, the discussion is between doing the barest minimum and doing something that in fact goes against a meaningful ethical principle.

While philosophical or ethical arguments are good over a beer, they rarely move the needle on policy of any sort as far as I can see, particularly when it comes to money. So, in what follows, I will try to make the argument in the form of how much of a financial difference there is between a low and a high wage earner when we adopt a proportional (regressive) salary increase scheme.

In what follows, I will present a series of calculations that show how extreme the results of a proportional raise scheme are. Note, on reading this, if you as a reader have any additions or corrections, please send them my way and I’ll update my post accordingly. I hope this post will be updated over time with more information and additional clarity.

Some simple calculations

Let’s imagine two faculty with the same merit ranking, so they should in principle deserve the same raise; however, one makes $70K (“low wage earner”) and the other makes $200K (“high wage earner”). Let’s also assume a regular percentage merit raise of 2% for those in this merit ranking.

In what follows, all calculations were done in R (embedded in Rmarkdown in this document) and rounded off in the text, so you might see some rounding off errors (but that should result in at most a single integer value difference in the last place of the number). The code is hidden to make the text more readable.

Difference in raise amounts for one year

The difference is actual raise amounts for the first year between the two faculty would be $4K - $1K = $3K.

Difference in salaries after 10 years

Compound interest has justly been called the “The Eighth Wonder of the World”⁴. To replicate the following calculations, a simple compound interest calculator can be used, if you’d like.

The difference in salaries at the end of 10 years would be $244K - $85K = $158K. This means that the high wage earner has an additional increase of $28K of salary in 10 years.

Difference in gross salaries over all 10 years

Now, one can compute the additional salary increase for each of the 10 years as in the table below (last column). Over 10 years, then, the low wage earner would have lost a total of $152K!

year	lowSalary	highSalary	diffSalary	additionalSalary
1	71400	204000	132600	2600
2	72828	208080	135252	5252
3	74285	212242	137957	7957
4	75770	216486	140716	10716
5	77286	220816	143530	13530
6	78831	225232	146401	16401
7	80408	229737	149329	19329
8	82016	234332	152316	22316
9	83656	239019	155363	25363
10	85330	243799	158469	28469

A more realistic difference

OK, people don’t usually save their money is a 0% interest rate account over 10 years. Instead, they are likely to save it in their 403b or some investment vehicle. So, what would the additional money have been if it were invested in a standard retirement plan? While most faculty tend to pick a Target Date fund, I will use a 60% stock+40% bond, all U.S. portfolio, to make my point — this portfolio is generally more conservative than the Target Date funds for most intended age periods of the latter fund, and therefore, the difference number I report is likely to be an underestimate.⁵

For the uninitiated, https://www.portfoliovisualizer.com/ is a wonderful website to plan your investments and retirements. In what follows, I will just use a long-term average annualised return for the period Jan 1987 - Jul 2025 as an estimate (=8.84%).

So, if the additional money were invested in a conservative 60% stock+40% bond, all U.S. portfolio, over 10 years, the low wage earner would have lost a total of $199K (sum of last column in the table below)! That is a massive amount lost. This is what is at stake.

year	additionalSalary	returnByEndOfPeriod
1	2600	5573
2	5252	10343
3	7957	14397
4	10716	17814
5	13530	20665
6	16401	23016
7	19329	24922
8	22316	26436
9	25363	27605
10	28469	28469

Of course, even if the low wage earner didn’t invest all the money, there is still the lost opportunities in spending the money on children, travel, housing, …

Some potential questions or follow-ups

What do I (Karthik) have to gain through this post?
Nothing. In fact, I’ll likely lose money if a flat or progressive raise scheme is implemented. I am being open about this, because as a public sector employee, I don’t believe I have any right to privacy regarding my salary. You are welcome to see my salary on the GovSalaries website. As a state employee my salary information should be publicly available; though MSU has made the information less transparent in recent years violating that basic principle of salary transparency by hiding the information behind a log-in entry. I plan to write about my own salary changes over the years and MSU’s violation of the principle of salary transparency in later posts.

Furthermore, I am declaring the lack of a benefit because I have a general rule of thumb — always take advice with a grain of salt if the person stands to gain from implementing the advice, not because people are bad, but because that’s life.

The salary that a person has is an aggregation of the merit raises of the last however many years, so a differnce in actual amount of raise in a single year is not unfair as it reflects merit over multiple years.
First of all, in my opinion, the largest contributors to salary gains are retention negotiations, and not merit raises. Furthermore, retention negotiations themselves in many places involve a lot of things that entrench inequity along multiple dimentions (White-Lewis et al. 2025), and are not reflective of merit alone (or at all?). Next, using retention negotiation related salaries as a basis for merit increases is extremely weird for a university/college/department as the individual has to go against the best interests of the university/college/department and think only for themselves. In fact, if retention related salaries are used as a basis for merit increases, the university/college/department is actively rewarding behaviour detrimental to the relevant unit. Here, I am simply stating the fact, and don’t mean to pass any moral judgement, as I myself have been part of such negotiations.

The merit raise pool of money given to a unit (college/department) is based on the total salary amount within that unit. So, a percentage raise is fair.
I don’t quite know the facts here, and whether this is true, but I believe it is irrelevant to the definition of merit. The argument that it should be a percentage of the salary because higher earners are bringing in more to the merit pool within a unit effectively says that the definition of “merit” should include the salary of the individual, and that is a definition most people will be uncomfortable with.

For similar reasons, I am also not in favour of a combination of flat+percentage either. Furthermore, it is still regressive and it is also more complicated.

A proportional raise is fair as it partially offsets inflation.
I don’t entirely follow this argument. Inflation affects all of our basic needs in raw numbers. The absolute cost increase in basic goods is the same for all of us. It is not the case that inflation forces a person getting $200K to pay more for basic goods than one earning $70K. For example, it is not the case that inflation causes a high wage earner to suddenly pay more for eggs, milk, bread, electricity, gas (and other things that politicians like to talk about) than a low wage earner. The price of basic goods increases by the same absolute amounts for all of us. So, unless higher wage earners for some reason need to consume more basic goods than low wage earners, we are all equally affected by it to the same absolute dollar amount. Of course, higher wage earners sometimes have higher amounts of discretionary spending (towards more expensive cars, houses,…), but this is their choice and can’t form the basis of inflation adjustment.

Conclusion

I have tried to show that, for two people earning $70K and $200K, an arguably conservative difference in net worth at the end of 10 years is about $199K. The difference would be $1.1 million in 20 years, and $3.9 million in 30 years. These are massive numbers, and this is the curse of compound interest on a low wage earner in regressive salary increase systems. Percentage-based salary raises are deeply regressive, and it is my strong opinion that they have no place in a society aiming to be fair.

In fact, proportional raises are inconsistent with MSU’s own policy in the case of promotions, where the amounts are flat raises, “at $4,000 per promotion from assistant to associate professor and at $5,000 per promotion from associate to professor”. The inconsistency with annual merit raises is both glaring and weird.

As I mentioned at the beginning, flat raises are the bare minimum. In a fair society, we would have progressive raises, where the lowest wage earners get the highest (dollar amount) raises. But, flat raises are a start on the path to fairness. Let’s start there.

I realise talking about salaries is seen as taboo, and somehow beneath academics, in academia. But, as with most silence, this silence only benefits those who already benefit from the system. I hope anyone who reads this post will use this as a starting point to discuss the issue openly with their colleagues and land on decisions that they collectively deem fair.

Outside the ivory tower, universities are often said to be liberal bastions. But the more I look around, the more I see quite regressive policies — policies that help the richer folks or higher admin. We live in times that have entrenched and expanded inequities, but we can’t complain about or fight such inequities, if we ourselves add to them.

---

1. You are able to get FOIA requests answered by MSU by emailing foia@msu.edu. See here for more on the issue.↩︎

2. It is an extremely poorly defined word, and can be quite harmful. It largely repackages historical contingencies and inequities as something useful. Furthermore, as many have written, merit-based systems are not good at achieving the ends they are set up to achieve — as they largely exacerbate inequities. But, I will play along here.↩︎

3. This also allows one to justify progressive taxes.↩︎

4. Not by Einstein, as it turns out. The earliest known reference to this was in an ad for The Equity Savings & Loan Company published in the “Cleveland Plain Dealer” of Ohio in 1925.↩︎

5. I say “likely” and not “surely” because the Target Date funds include Total World stock and bond funds, while I just used U.S. stocks and bonds, and so one can’t claim to be sure.↩︎