Our Compass May Be Off

by Michael L. McCune

Before I begin this article I must first admit to something in my dark past- I was an internal auditor for a couple of years when I got out of college; until I found a real job! However, old habits die hard and to this day I have retained a compulsion about reconciling and understanding what is behind numbers. In my short history as an accountant I learned that there is a major difference between "actual", "projected" and "statistical" numbers.

For several years I have been an avid reader and user of Self-Storage Almanac. Much of the information is interesting and useful. However, over the years I have found what I consider inconsistencies in some of the data relating to total self storage square footage (nationally and locally) and saturation levels. I advice my clients and our friends to be aware that statistically derived (using certain assumptions) and not necessarily totally based upon empirical data. Thus, on a recent long flight I decided to take another good look at the 1999 Almanac and see, if indeed, the reported statistical data meets my "reasonableness test" and, if not, what are the potential ramifications. I have presented my analysis below and you can be the judge as to whether the statistical data meets your "reasonableness" standard. I have developed some opinions as to what the ramifications might be if the statistics are not representative of the real world. See if you agree.

The methodology used in the 1999 edition to develop the total square footages was to survey about 2895 self-storage facilities with 472 responses, obtaining facility sizes and then averaging them. That average size (38,650 sf.) is then apparently applied to a data base of all self-storage facilities (27,535) yielding a total self storage square footage of about 1,064,000,000 square feet. This all seems quite reasonable at first, except when you realize that for every 10,000 sf. facility there must be a 66,650 sf. facility in order to maintain the average facility size of 38,650 sf. Similarly, for every 5,000 sf. facility, there must be a 71,650 sf. facility to sustain this average. My intuition tells me that across this great land there are many more small facilities than large ones and that the 38,650 sf. average may well be too high. I don’t know what the average size is, but I think 38,650 sf. may be too large for an average of all facilities.

Pursuing this notion a little further, let’s examine a market a little closer and see if the numbers make sense in the context of the cities and states:

Let’s start with my home state, Colorado, and take a look at how these numbers work out. First, we know from our actual count that the Denver area has about 8.4 million square feet, so lets just say the Almanac’s 7.2 million is close enough given that different definitions of area may exist. But then a little math takes us to a different place in Colorado, namely, rural Colorado. With all due respect to two or three other great Colorado cities, there is not much else in Colorado that qualifies as a metropolitan area other than Denver and Colorado Springs. Yet using the Almanac data and "backing" into the rural Colorado’s numbers would indicate that rural Colorado utilizes self-storage at a rate of almost 2.5 times as high as the Denver Metropolitan Areas on a per capita basis (i.e. 10.21 sf./person in rural Colorado vs. 4.01 sf./person in the Metro areas). Having had an opportunity (in Colorado we say privilege) to drive through much of this area and with more than a casual interest observing self-storage, I think my anecdotal experience would be that there aren’t that many large facilities (certainly not 406 facilities averaging 38650 sf. each) out there.

I think that if you take a look at the percentage numbers of population related to the percentage number of self-storage square feet per the Almanac, the numbers may not fit your perception of the empirical world either. These numbers indicate that, despite a smaller population, a rural Coloradoan would use more self-storage than a person in urban Colorado would. In other words, the data indicates that the rural 38.8% of the population use 61.8% of the self-storage. It doesn’t make sense to me that if you own a barn you might rent more self-storage than your city cousin would. My intuition also tells me that rural areas have fewer apartments and less industry than Metropolitan areas, both of which we believe are significant users of self-storage.

I have performed the same type of analysis on Arkansas, Louisiana, Oklahoma, South Carolina, North Carolina, Utah, and Arizona, and found a very similar pattern of results although the magnitudes varied. Those states were the easy ones to analyze because of the dominance of one or two large cities, but I suspect that this pattern would apply to many other areas. My suspicion is that the perceived distortion may even be more pronounced in the east where "rural" zoning is more difficult, but there isn’t enough information to develop clear examples.

I also ran some numbers on computation of the Supply Index on pages 93&94 of the 1999 Almanac, which is designed to show a percentage deviation by city from the total average, i.e. the square feet/person for a given city divided by the total average square feet per person for all the cities. It is interesting that the index for the largest 50 cities is computed based on a total average of 2.87 sf./person and 3.97 sf./person average for the next 50 largest cities. Since the denominators aren’t the same for the two lists, it appears that the Supply Index for the two lists may not be mathematically consistent or meaningful when comparing cities on the different lists, (i.e. Middlesex, NJ’s Index (the last item on first list) can’t be compared to Hartford’s Index (first on second list)). For example, if Hartford had been on the first list, it’s Index would have been 96.2 rather than 69.6. The only apparent difference is that Hartford was on the second list because it had seven thousand less people than Middlesex.

It is also interesting that, according to my simple division of the statistics on pages 91 through 94, the average size of facilities in every state and every city described in the Almanac exactly 38,650 square feet and that 38,650 sf. is also the national average. Not impossible, but it seems to be a highly unlikely coincidence.

You may be saying "Oh, so what! You wasted your flight, what does it mean to me?" While I can’t quantify the potential ramifications, I think they may be apparent from my reasonableness tests. My ideas and opinions on these ramification are listed below:

• The total amount of square footage of self-storage may be overstated in the Almanac, perhaps by a fairly meaningful percentage. This could mean that the concentration by the large owners is greater than previously thought. Thus, these owners may have greater market influence than previously thought.
• It could also mean that reliance on the number of square feet per person in a given city or state based on the Almanac data might produce a decision not consistent with the actual empirical market place. Thus, a potential developer might believe that the square feet per person in his market is 6 square feet and find an area where he empirically calculates only 4 square feet and thinks he has a good market, only to find out later that the Almanac numbers didn’t represent the real empirical market, but rather, as disclosed, a statistical average applied to a city or state. The chance of overbuilding may be increase, thus potentially hurting all market participants.
• Distortions in the supply index computation methods could cause some cities to either be overlooked or too highly evaluated in some analyses.
• It is clear that if confusion in these numbers exists, it should be the highest priority of our industry to get a definitive measure and actual empirical data of the total self-storage market, by metropolitan area, state, and national area.
• Other tables and charts also may be based on sampling methods and projections and should be evaluated accordingly when you are making business decisions.

In the meantime, I advise our clients and our friends to be aware that the square footage totals in the Almanac are statistically derived and are not necessarily based upon empirical data.

While I think that it may be difficult to criticize others’ work, especially work created in good faith as the Almanac is, I feel that it is important for all of us in the industry to try to and have the best perspective on the information available so that we can properly evaluate our alternatives and the information available. As the industry grows, data must be better supported by empirical as well as statistical details.