I recently got really into puerh and spent some time thinking about the relationship between age and price. The common wisdom is that puerh tea will increase in price every year it ages, with a significant increase at 4-5 years. White2Tea even warns its customers of annual price increases. However, while there is a lot of folk wisdom about puerh prices, there is very little data. So, as I have done in the past, I decided to collect my own.*
*This post has some technical language, but I do my best to explain what I mean by technical terms. There is also a short appendix at the end.
Take a look at the following figure:
After collecting a random sample of 230 puerh teas offered by Yunnan Sourcing (as of August 2017), I graphed the relationship between price (y-axis) and age (x-axis) for sheng and shou puerh. First, notice that the relationship between price (per 100g) and age is positive (corr. = 0.37) and, if you run an OLS regression, significant. In fact, each additional year of age is associated with an average increase of $1.28/100g (p < 0.001), when controlling for style (Sheng = 1) and shape (Cake = 1) (Table 1). Second, looking at the graph, there is a noticeable increase in price around the tenth year. These trends are clearer still when you look only at sheng costing less than $50/100g:
In this subset of the sample, you can see a clear increase in puerh prices over the past 19 years (1999-2017). Moreover, at the lowest price range, there are obvious increases at years seven, nine, thirteen, and sixteen—or about every three years. (You can even see higher prices for 2008 puerh, perhaps a legacy of the bubble.) With a larger sample (and many more vendors), I bet that these trends would stand out more, though the trendline would smooth slightly. Another interesting trend is the small annual increase in sheng price for teas produced in 2013-2017. In those years, the relationship between price and age is actually negative. My guess is that puerh offerings (esp. the Yunnan Sourcing label) have become much higher quality since 2013. I think that’s probably a good thing.
The positive relationship between price and age also emerges when looking at regular productions. For example, the 2017 Menghai 7542 sheng costs $44, while the 2015 production costs $47. The Yunnan Sourcing label “Impressions” sheng costs $23 for the 2015 production, and $29 each for 2012 and 2014. At the higher end, the Yunnan Sourcing label “Mu Shu Cha Ancient Arbor“ sheng costs $120 for the 2017 production, $144 for 2014, $152 for 2012, and $154 for 2011. For shou, the 2017 Menghai 7572 shou costs $25, while the 2016 production costs $27, 2014 costs $29, 2012 costs $33, 2010 costs $35, and 2009 costs $38. For these four productions, the annual price increase is about $1.50-2.00 per cake at the lower end and $4-7 per cake at the higher end.
Compare those trends with White2Tea:
White2Tea has the opposite relationship. The quality (and price) of high-end W2T teas have increased so much over the past few years that they’ve actually outpaced the price of age. Hence, the trendline on the figure above is negative, as is the price-age correlation (-0.20), though it’s not significant (Table 1).
What about YS and W2T “Label” teas?
Finally, looking only at YS Label and W2T Label puerh you can really see the differences in price and price/age. First, it’s worth noting that Yunnan Sourcing has been producing private label puerh since 2009. White2Tea, however, has only been producing since 2012. (I did not include W2T teas before 2014.) Second, the average price of W2T Label puerh is a little over two times higher than YS Label puerh ($47.89/100g versus $22.31/100g). Again, the association between price (per 100g) and age is positive (but not significant) for YS Label puerh and negative (but not significant) for W2T Label puerh.
These two companies provide a nice snapshot of the current (western-facing) puerh market more generally. Large-scale vendors like Yunnan Sourcing (perhaps especially Yunnan Sourcing) have a wider range of products (beyond their own label) at wider price points. On the other hand, there are smaller specialty vendors like White2Tea (or, for example, Crimson Lotus Tea) who primarily sell their private label teas and may have higher average quality. I think that this sort of bifurcation is actually good for consumers and is a necessary component of a robust puerh economy. However, if the price difference between “budget” and "specialty” vendors gets too large, then it will inevitably lead to a bifurcated customer base (and tea community). This has not happened yet, though the trends for W2T may foreshadow a future shift.
To create the sample of Yunnan Sourcing puerh, I collected data on every puerh offered online by YS (sorted by age) produced 1990-2003 and then every third puerh thereafter. Teas that were “out of stock” were not included in the sample, unless they still had prices listed. Only pressings of 100g or higher were included. I collected data on price ($), weight (g), age (years), type, shape, and label. I did the same data collection for White2Tea puerh, except I included all puerh (over 100g) in their online store for which prices were listed. I do not include W2T “label” productions before 2012. Because the teas labeled 2012 and 2013 were both out of stock, the first production included was from 2014.
For analysis, raw prices (per cake) were converted into price per 100g and year was converted into age (2018-year). To check for robustness, analysis was repeated with price per cake and price per 100g. The latter was chosen for ease of interpretation. Controls for type of tea (Sheng = 1) and shape of pressing (Cake = 1) were also included in the multivariate regressions. While more data—both a larger N and more variables—would be useful for further analysis, it is not practical in this context. Moreover, limiting analysis to one large vendor (Yunnan Sourcing) and one specialty vendor (White2Tea) provides an interesting comparison, while limiting other effects (e.g., storage, shipping).
Below is a regression table showing the relationship between price (per 100g) and age (years), with controls for type of tea (Sheng = 1) and shape of pressing (Cake = 1). The analysis shows that among puerhs sold by Yunnan Sourcing, there is a positive and significant relationship between price and age (p < 0.001). Within that same sample, sheng puerh sells for about $8.80/100g more than shou (p < 0.01). There appears to be no effect of puerh shape, though that might be because sheng is unlikely to be pressed as anything but a cake.
Among YS Label puerh, there is no significant relationship between price and age, although that may be because the sample size is small (N = 82). The same applies for the two control variables. For W2T Label puerh (N = 61), the relationship between price and age is negative but not significant. Not surprisingly, there is a significant difference in the price of sheng versus shou, with W2T sheng costing about $37.98/100g more than W2T shou (p < 0.05).
(If you have any questions about the data analysis, or have suggestions for future data collection projects, please let me know.)
