Thursday, October 16, 2008

Will the economy hurt drug stores?

The economic outlook for 2009 looks grim. The Wall Street Journal’s October survey of economic forecasts puts the odds of recession at 89% and forecast negative GDP growth through the first part of 2009. U.S. retail sales dropped sharply in September, leading to another down day on the stock market.

So what, if anything, does this mean for overall sales at drug stores in 2009?

Based on the historical evidence … not much.

I looked at year-over-year changes in quarterly retail sales at Pharmacies and Drug Stores (NAICS 44611) as collected by the Bureau of the Census. These government-collected data provide the most complete picture of revenues at all drug stores, not just the big public companies. Note that these data exclude non-pharmacy dependent retail formats, such as supermarkets or mass merchants, and also combine prescription and front-end sales.

I compared these data to U.S. Gross Domestic Product (GDP) and Total U.S. Retail Sales data. The chart below tells the story – there was no clear relationship between changes in GDP and drug store sales during the past 8 years. (Click to enlarge the chart.) There was also no clear relationship in the 1990s (not shown).

Changes in sales at pharmacies and drug stores were not strongly correlated with changes in either GDP (-40%) or total retail sales (-41%; not shown). In fact, the correlation was actually negative, i.e., drug store sales tend to go up when GDP goes down. In case you’re wondering, changes in total retail sales had a +90% correlation with changes in GDP over the same time period.

The chart above also illuminates the sharp slowdown in year-over-year growth at drug stores that began in early 2007, which was well before the recession started.
I think this trend reflects more fundamental industry forces at work, such as a lack of new blockbusters and rising generic utilization rates (which reduces revenue but boosts profits). The nature of health care products also implies limited sensitivity to a downturn.

further piece of evidence comes from same-store sales at major retailers. (See Consolidated Summary of Sept Retail Sales Results.) Twenty-eight of 36 major retailers had negative same-store sales growth in September. Who was up?
  • Discounters: Wal-Mart (WMT), Costco (COST), BJ's Wholesale (BJ)
  • Drug Stores: Walgreens (WAG), Rite-Aid (RAD), Fred's (FRED)

Do these facts “prove” that drug store revenues will hold up in 2009? No, of course not. The future could look quite different than the past.

But as Mark Twain said: “History does not repeat itself. But it rhymes.”


  1. Great analysis, Adam. But your last comment about discount stores makes me wonder about the impact of $4 generics.

    If discount stores are gaining during a recession AND they are focused on pharmacy now, does it mean that drug store sales will be lower in the future because they lost market share?

    I'm thinking about your comments on Wal-Mart a few weeks ago and wondering if the recession will make them much bigger in pharmacy by 2010 or 2011. Any ideas?

  2. Sounds like you disagree with Jeff Rein!

  3. Regarding the second comment above, today's AIS Health Business Daily newsletter leads off with the following quote:

    "We are facing a continuation of the slowest-growing prescription drug market in 47 years, according to IMS Health. We believe the biggest impact has been the very tough economy." — Jeff Rein, chairman and CEO of Walgreen Co. Source

    An especially interesting statement in light of Mr. Rein's recent "retirement."


  4. Adam,

    Using Bureau of the Census data, I calculated the correlation between total retail sales and pharmacy sales back to 1992. My result is 0.99, far far away from -0.40 mentioned in your blog. I wonder from what data did you get the -0.40 number. Thanks for your time.

  5. Re: the correlation

    I correlated the *change* in sales at pharmacies to the *change* in retail sales. I believe that you are reporting the correlation between the *level* of sales at pharmacies to the *level* of retail sales.

    If you correlate the level of sales, then the correlation is artificially inflated because there is a common time trend a.k.a. a unit root.

    In other words, retail sales of all kinds tend to grow over time, so correlations of sales levels are really just showing the correlation of the time trend. Taking the first difference (change in sales) removes the time trend and highlights the true relationship.

    Sorry if that was not clear. I edited the text of the post to clarify that I am discussing correlation between the changes.

    Note that I also aggregated the Census data to calendar quarters to smooth out the monthly variability.


  6. Adam,

    Thank you for your explaination. Using the "change", the correlation is -0.1. Not quite at -0.4, but the merit of your argument is well-taken. Thanks.

  7. It is great analysis. You analyzed the reason about the sharp slowdown in year-over-year growth at drug store. Furthermore, Could you give us a general explanation about the trough for drug store revenue happened during 2004 and 2005?