In September, I’m heading to Canterbury , New Zealand as part of a joint project between the World Bank and EERI . The objective of the project is to research how and whether recovery after the February 2011 earthquake reflects the concept of “build back better.” Apparently, the World Bank is keen on developing guidelines that are relevant across different types of disasters and countries for facilitating building back better.
Being the community resilience nerd I am, booking my travel made me wonder how tourism is doing in Canterbury and Christchurch.
To be able to answer this, first, “recovery” needs to be defined. This is, of course, an issue that the field of community disaster resilience continues to struggle with and one of the reasons there is lively debate, as well, about the definition of resilience itself.
What is the baseline for recovery? How do you know when recovery has occurred? Is recovery about speed? Or is it about quality? Is recovery to some pre-disaster state or an alternate reality state where there was no disaster or some new normal altogether?
So many questions and, frankly, these questions are posed more often than concrete proposed answers are given. I don’t have THE answer of course, but I think I have a significant contribution to the discussion. And I think the question about tourism in Canterbury is a very good case to illustrate my thinking.
In most cases, we see recovery defined or measured based on the supply of something–the supply of housing or water or social services, and the like. Part of the reason why is that supply matters. Another part is that supply is relatively straightforward (though not always easy) to measure and get data for. These two reasons are exemplified by housing reconstruction is a focus of recovery research.
Typically, evaluation of the recovery of the supply-side of various things is based on speed–how quickly has, for example, the housing stock returned? Was housing reconstruction as efficient as possible?
But as Rob Olshansky has brought up in , there can be a tension between speed and quality of recovery. Sometimes you have to go slow to go fast. Sometimes you need to be deliberative in order to ensure everyone has a chance to recover and is maybe even better off.
Now imagine I’m administering an implicit association test to you and say the word “supply.” What would you say? I’m going to guess: “demand.”
Supply and demand!
The idea of demand recovery is much less common in the literature. But clearly, without demand recovery there can be no, say, business recovery. A business might have their building back and functional after a storm. But, if there are no customers, there is no recovery. I mean, how many people were patronizing high-end pet boutiques in the weeks after Hurricane Sandy? Probably fewer than in the weeks before. If there is not huge demand, what’s the rush in recovering supply?
So understanding recovery and, so, resilience requires consideration of both supply and demand. When you consider both supply and demand, you see recovery as relative. Recovery then relates to the balance of supply and demand. When I think about supply and demand, I no longer think about just efficiency; I also think about sufficiency.
Oh yeah: the recovery of tourism in Canterbury.
Being the geek I am, I had to see if there was public data that I could use to get a sense about the sufficiency of tourism recovery in Canterbury. God bless Statistics New Zealand who have the amazing data tool Infoshare that made tracking down interesting time series data very simple and fruitful.
To illustrate my thoughts, I grabbed [xlsx] about various hotel indicators that we’ll use as a proxy for tourism: number of hotels (hotel units), number of nights people stay (guest nights) for both international and domestic travelers, and the relative use of the hotel stock (occupancy rate). In the following five graphs, I’ve plotted these indicators in raw form, in addition to a centered moving average, additive seasonal adjustment, and multiplicative seasonal adjustment.
There are definitely fewer hotels after the earthquake.
It seems like there is less use of those hotels too.
Though, if you look at just domestic guest nights…
…you realize that, really, its just international guest nights that have gone down.
Interestingly, the average occupancy rates of the existing hotels is pretty similar to before the quake, maybe a bit higher depending on how far back you compare.
But is the supply coming back efficiently? I’m too lazy to actually compare trend lines, but the rate of growth seems about the same, both before and after. Though maybe you think growth should be faster after than before?
Looking at the above five charts, it is rather obvious that from a supply standpoint the tourism industry (vis-a-vis hotels) has not returned to its pre-disaster state.
What about demand recovery? This is a bit more complicated to answer. After versus before, there are fewer guest nights, overall, but the occupancy rate is about the same. If I’m understanding the indicators right, this is saying that at any given hotel today you will see about the same number of people at that hotel, but there are fewer people booking hotels and/or people staying fewer nights per trip. Is that right?
As I said before, these questions independently are not as useful as coupling them: is the supply-demand recovery curve trending up or down?
To get some insight into that question, I created the graph at the top of this post. To make it, I first calculated “available units.” To do this, I multiplied the hotel units by occupancy rate to figure out how many hotels are fully occupied. (Yes, I realize that’s not how the supply and demand of the hotel biz works, but lets pretend it does.) I then subtracted this product from hotel units to get an idea of the number of available hotels. Lastly, I took the inverse of the result and multiplied it by guest nights. (That is, I divided guest nights by available units.) The final result is a rate that normalizes the total number of guest nights in Canterbury hotels (demand) by the number of unoccupied hotels/rooms (supply).
And what do we see?
Arguably, we see that the relative overall relationship between supply and demand has improved since the earthquake. That is, given the number of hotel units left, their use has gone up. Given the hotel units in existence, the hotel industry seems to have recovered because supply is meeting demand. This is perhaps not surprising if you think about it, which is born out looking at the different trends between domestic travelers and international travelers. Its the domestic demand that has gone up. This is probably the result of New Zealanders coming to Canterbury to take new jobs associated with reconstruction and recovery.
The February 2011 earthquake had a clear impact on the sufficiency of the hotel supply with respect to domestic travelers. But what about for international travelers?
The earthquake had no effect on the supply-demand relationship of hotels with respect to international travelers. International tourism has recovered! Well, at least based on the idea of sufficiency of recovery.
In fact, the supply-demand balance is roughly the same as before the earthquake. Before the earthquake, Canterbury was building hotels faster than they were generating demand for them, particularly international. Now after the earthquake, the demand has gone down, but the supply went down faster.
Curious how I did the analysis and assembled the graphs? The code is and the data is [xlsx].