Supply Curves and Cycles
While most goods and services have an upward sloping supply curve, there are several informative exceptions to this pattern. This post will explore the shape of the supply curve of four different assets, bicycles, residential real estate in Las Vegas, residential real estate in New York City, and bitcoin.
Bicycles are widely understood to be a normal good produced in a competitive market. The supply curve for bicycles is upward slowing and elastic. For any given shift in demand, producers can quickly increase or decrease production in the short-term.
Unlike bicycles, the supply curve of residential real estate is kinked at the origin. The supply curve for housing has a kink at the existing level of housing because housing is durable and does not diminish quickly when demand falls. (Glaeser and Gyourko, 2005) The slope of the supply curve for residential real estate to the right of the origin is related to the ease with which developers can increase the housing supply in the short to medium run. Building in New York is far more difficult due to land scarcity, zoning, and regulatory complexity among other factors relative to Las Vegas.
Bitcoin’s supply curve is among the most inelastic of all known assets. This is because the supply schedule of bitcoin is fixed. Even if additional miners enter the market, the mining difficulty will increase, slowing the subsequent rate of new production.
Supply Curves (Red), Demand Curves (Black)
If we assume an equivalent rightward shift in the demand curves for the four assets depicted in the chart above, we can conclude that the steeper the supply curve, the greater the expected increase in price.
Supply Curves (Red), Demand Curves (Black)
If we assume an equivalent leftward shift in the demand curves for the four assets depicted in the chart above, we can conclude that the steeper the supply curve, the greater the expected decrease in price.
As is clear from the above two charts, assuming an equivalently sized shift in demand to the right, bitcoin’s price volatility is expected to be the largest of the four assets. Assuming an equivalently sized shift in demand to the left, bicycles exhibit the smallest price decline. This is because bicycle manufacturers can easily cut production to limit price declines. On the other hand, bitcoin and real estate should experience comparable price declines that are far steeper than that of bicycles. Unless cities resort to large-scale demolition programs, real estate supply persists through time even if demand falls. Like real estate, bitcoins persist through time. In fact, the only way for bitcoin’s supply curve to shift to the left is for private keys to be permanently lost.
Unlike real estate and bicycles, we also know the shape of Bitcoin’s long-run supply curve. The number of bitcoins will continue to increase at a decreasing rate until such time that the number of bitcoins reaches 21 million around the year 2140. In the early years, bitcoin’s supply curve was moving to the right rapidly, but at this point, only about 900 newly mined bitcoins are produced each day (26 million USD at current prices).
Cycles are caused by both changes in fundamentals and changes in expectations about future fundamentals. In the case of bitcoin, supply fundamentals and expectations about future supply fundamentals are both known with a high degree of accuracy and precision. Therefore, all future price changes will be caused by changes in demand fundamentals and expectations.
Like real estate and public equities, bitcoin markets have both short-term traders and long-term investors. As the proportion of short-term traders grows, the higher the probability that there will be a reduction in prices, once these traders decide to “take profits” or borrow to engage in short selling.
Unchained Capital, using data from bitcoin’s blockchain charted what they call HODL Waves. As the percentage of bitcoin that has transacted recently increases relative to the share that has been dormant for long periods, we should be wary of a bubble.
The colored bands show the relative fraction of Bitcoin in existence that was last transacted within the time window indicated in the legend. The bottom, warmer colors (reds, oranges) represent Bitcoin transacting very recently while the top, cooler colors (greens, blues) represent Bitcoin that hasn’t transacted in a long time. The chart has been normalized by the BTC in existence at each date (left y-axis). The black line shows the USD/BTC price (logarithmically, right y-axis).
As is evident in the far right of the chart above, it looks like another bubble is in the early stages of forming. However, we don’t know whether recent buyers are likely to hold for many years or sell in a few months. There is some reason to suggest that the composition of buyers these days may be more long term oriented relative to the previous cycle.
Nonetheless, Google Trends data and Twitter Trends data below also suggests that hype is growing materially.
Bitcoin is about 31% as searched relative to its peak search interest in late 2017. The data suggest a clear upward trend.
Bitcoin is about 50% of its Tweet frequency as the peak of the previous cycle. The data suggest a clear upward trend.
Overall, we have strong theoretical reasons to believe that bitcoin prices will be cyclical over time. However, if we expect high price volatility and cyclicality, perhaps our expectation of the events will serve to attenuate them. I believe we are entering the early stages of a cycle which will likely lead to a meaningful increase in price. However, along with this belief, I understand that a 50% decline in price is entirely possible and well within expectations. I’ll be watching the Google trends, Twitter trends, and HODL Wave data closely.
“The secret to happiness is having low expectations.”