Efficient Market Hypothesis (EMT) is the Nobel Price winning research was developed by Prof. Eugene Fama along with Prof. Kenneth French in relation to asset pricing. The hypothesis, to paraphrase, states that the price of the asset at any point of time reflects any information available on the asset at that time. A direct conclusion of this hypothesis is that it is impossible to “beat the market” or in other words, break away from the underlying market structure and economic forces affecting the demand of your asset. The hypothesis has had both supporters and critics over the years since its first publication in 1998. Since then, the hypothesis has been found to work best in low volatility conditions and when we look at asset prices over longer periods.
However, the hypothesis is related to “asset” pricing and it does lead one to wonder if the logic embedded in the argument holds true for the hospitality industry as well. At any point the most important information available with respect to a hotel’s asset, which in this case is rooms available for a particular night, is its demand for the same room. The intuition as to how EMT and demand could be applicable to hotel’s room pricing is simple: If the market price of the room is lower than what available information (demand) would suggest it should be, customer could (and would) profit by buying or renting the asset (in this case, the hotel room). This increase in demand, however, would push up the price of the asset (available room) in efficient markets, until it was no longer underpriced or exceeds its price-value proposition. So in the below figure, due to increased demand, the room’s price moved from market in-efficient P1 to an efficient P2 where it matched the demand.
However, in order to apply the principles of EMT and by association, basic micro-economics of demand and supply, we need to determine the demand for the asset in question. Therefore, in efficient markets this demand will then manifest in the pricing of say a basket of rooms in a given market. This brings us to the area of demand forecasting within revenue management and its intrinsic correlation with the market-driven efficient price of the available room. There are existing models for demand forecasting within the revenue management domain ranging from long term ARIMA models based on annual cyclicality in the hospitality business which perform well for medium and long term demand to Bayesian models and additive pickup models which perform really well for short-term demand. The main question is this: can we judge and assign the market-efficient price for any room type of a property by simply looking at forecasted demand for the same. This is an interesting way to look at efficient pricing of the hotel’s assets using market forces instead of a factor based approach which estimates room price based on external factors such as the city, nearby competitor hotels or tourist traffic coming into the city.
EMT would suggest that such information is already factored into the demand for the room for that night, so we could use the demand as the sole guiding factor for pricing determination. However there are times that the inherent demand is not captured in on the books data and is typically driven by Black Swan or one-off events. Examples could be an event that is not yet known to the market but could be manifested in other signals such as airfares, vacation rentals etc. At Hotelsoft, we capture similar data sets to get early indicators of demand volatility. Hotelsoft RMS therefore has a combination pricing model that utilizes demand based pricing along with the third party signals to come up with the best pricing for the hotel.
Building the system involved bringing in some real world aspects into the problem: many hotels prefer their pricing to move in steps, i.e, they do not want any price for their rooms to move in, say, less than $25 steps. Another constraint that we brought in was that each hotel has a comfort zone of price ranges they can work with in a given season, also known as benchmark rates, say from $200 to $700 for their standard room on any date in a given period. Our system needed to honor that range at all times.
However, even we kept the system bound between such constraints, the results were quite surprising, in a positive way. Our calculations of market efficient pricing matched where the price should be in all cases. Our system not only took cognizance of the current occupancy for the future stay date in question, but was able to the price of the room in relation to the changes in future demand for the room in question. Rooms with high demand had steadily increasing prices as forecasts started including higher pickup. Revenue Manager is also able to incorporate data into our system that is not yet captured in the demand, thus getting best of the two worlds.
