Measuring Purchasing Power in Biomedical Research


How can we at NIH compare a $1 million investment in research today with the same $1 million investment twenty years ago, ten years ago, or even five years into the future?

You might be familiar with the Consumer Price Index (CPI), which measures inflation experienced by consumers in their day-to-day living expenses, or the Gross Domestic Product (GDP) Price Index, which measures general inflation across the nation’s economy. But as readers with an active interest in biomedical research, you might be interested in how these numbers ‘translate’ into the research world. Naturally, so are we, and this is where the Biomedical Research and Development Price Index, or BRDPI, comes in. Every year, NIH engages the U.S. Department of Commerce’s Bureau of Economic Analysis (BEA) to calculate and model the estimated changes in the costs of NIH-funded biomedical research. The BEA uses data from extramural institutions as well as the NIH intramural program to assess changes in the cost of items including personnel services, various supplies, and equipment involved in biomedical research. This information is used to calculate BRDPI as described in the annual BRDPI memo.

Actual and estimated inflation rates are posted on the NIH Office of Budget website, and these data were recently updated for 2013. I thought it would be interesting to share how both the BRDPI and GDP Price Index have changed from 1980 to the present, as well as price index projections for future years.

Graph of BRDPI and GDP Price Indexes from 1980 -2013, and projections for 2014-2019. Line at FY2012 indicates salary cap reduction of principal investigators. For source data: U.S. Commerce Department, Bureau of Economic Analysis, FY 1980 to FY 2019; OMB FY 2015 Economic Assumptions; NIH Office of Budget ( For description visit

You can see that the preliminary estimate for FY 2013 is 1.9 percent, up from 1.3 percent in FY 2012, meaning that to purchase the same level of research in FY 2013 as we did in FY 2012, NIH would have needed to increase its budget by 1.9 percent. The 2013 BRDPI is an estimate because it will be updated next year as global price indexes and other factors used to estimate the BRDPI are updated. You might also notice that in FY2012, BRDPI reached its lowest since 1980, which was primarily driven by the reduction in the salary cap for principal investigators.

Tracking this price index is important because it allows us to measure changes to the purchasing power of the NIH dollar over time, and make projections for future fiscal years. As described in the annual BRDPI memo, forecasting a future path of price changes is an inherently imprecise exercise; however, these forecasts are beneficial to our agency’s overall planning efforts by allowing us to consider the future value of a dollar specific to the goods and services used to support NIH research and development.


  1. I am not sure how the index is calculated — one way would be to compare the cost of doing a specific experiment now versus 5 or 10 years ago — but more realistically would be to consider the cost of getting a publishable result in answer to a specific question now compared to 5 or 10 years ago. The reviewing world expects us to use the most up to date procedures to address questions – often requiring use of high tech and expensive equipment or assays ( eg an epifluorescent microscope vs confocal or super resolution imaging; microarray vs RNAseq etc). If these technical advances are not included in estimates of increased cost, than the estimates are meaningless.

  2. This plus no change in the CAP for RO1 modular budgets in quite some time. Also salaries for students and postdocs (we are asked to follow the guidelines of the NIH here) and staff have gone up. And would be better to use alterations in costs for scientific supplies/equipment, hard to do I know due to the diversity of the research enterprise. All in all, I would bet that a grant at the modular cap with the same staff/student/postdoc commitment has 60% as much real $ to use for supplies/equipment animals etc.

  3. It is clear from these data that grants awarded in the mid 80’s for example are at about the same level as they are today. After grant award, they are further cut by 20% so if you have grants at the cap they are worth about $220K/yr and when compared for purchasing power they are worth about $100K compared to 1985. Keeping in mind that students and post-docs were considerably cheaper (PD were often paid less than $10K in the early 80’s while today they are about $45K to begin with). So not only are grants being funded at low rates but the funds are woefully inadequate. Given that most institutions are demanding large salary components (at Pitt it is 75% of the cap) we would need at least 2 R01 grants to do the work of one in 1985. This is unsustainable.

    1. I have already plotted these data using 2001 as a baseline, when the 250K modular was instituted. In 2013 dollars, one would need roughly 375K annually. I believe that reviewers are judging the scope of 5 year R01s given this level of funding, so, for a NI/ESI like myself, where a large chunk of PI salary and some new equipment needs to be on my first R01, going modular seems like setting oneself up for failure. In terms of the spending power, the current 250K buys nearly 175K would have in 2001. I think plotting annual percentage change is misleadingly rosy.

  4. Just from quickly glancing at the figure, I would have to conclude that this index (BRDPI) is not even close to being an accurate indicator of how the cost of doing research in a typical academic laboratory is increasing. Perhaps if one is committed to simply doing the same outdated procedures over and over again, and no additional costs are imposed by internal and external administrators, this index could give a rough idea of how well funding is keeping up with inflation. However, those are not the conditions we are facing. We must continually evolve to using newer and almost invariably more expensive methodologies, and with increasing competition must use these in greater and greater volume in order to outstrip competitors and avoid professional extinction. We are just now facing the prospect of a substantial increase in postdoc stipends; this is easily justified on the grounds that postdocs are poorly paid, but NIH and institutional administrators are forcing this on us without providing any funds to pay for it. The system is broken, and I fear its near complete collapse is imminent. It is really a shame, because the academic research enterprise in the USA was until fairly recently something that we could truly point to with pride. Not any more.

  5. Are these data annual increases? That is, are they calculated from the year before?
    If so, then we need to know the cumulative effect of 5, 10 and 15 year trends. In addition this needs to be compared to what the NIH budget was each year, so we can determine the real spending power of the NIH. Also, how do these numbers effect the number of R01 awarded? Lastly, it would be nice to know how these numbers influence the success rate, which has decreased recently. Could this information be helpful in determining what is needed to restore, maintain, and possibly improve the success rate to at least 25%.

  6. I wish to echo the comment from Erick above about the full modular grant limit to $250K direct. I had an analysis of this in 2012 on the blog, but really it isn’t that complicated to adjust a dollar number against some inflation index. They all tell a relatively similar story, whether we insist that competitiveresearch costs run ahead of standard inflation or not. Either way, the 1-2% described as needed between two fiscal years in this post is unbelievably misleading.

    Let us also keep in mind that 10% whacks to the full-modular award are a given these days.

    The Rock Talk blog spent a fair bit of time analyzing the grant submission churn, success rates, number of grants per PI and related data over the past few years.
    Somehow they never seem to recognize,in these analyses that purchasing power and relative funding needed to sustain a lab’s work is important. When I sit down and sketch out a research program approximately equivalent in scope to what I was writing in the early 2000s (for the modular limit) and do the traditional budgeting, gee, I come up with about $375-$400K in direct costs. Amazing how this reality matches the inflation indices, isn’t it? I’m sure other PIs have similar tales to tell….

  7. My experiences mirror Drug Monkey’s. I just ran the analysis on my cost sheets. I started my lab in 2006. My costs in personnel, mouse costs, reagents and supplies (in addition to increased requirements for my own salary coverage) put the increase at 42% per person over the eight year period from 2006 to 2014.

    My R01 renewal (started March, 2014) had a cut from $250K approved to $205K. This is the amount it was initially funded for in 2008 and in no way keeps up with the cost of inflation.

Before submitting your comment, please review our blog comment policies.

Leave a Reply

Your email address will not be published. Required fields are marked *