By Hubert Horan, who has 40 years of experience in the management and regulation of transportation companies (primarily airlines). Horan has no financial links with any urban car service industry competitors, investors or regulators, or any firms that work on behalf of industry participants
Previous Cfdtrade Uber analysis now published as a law journal article
My Cfdtrade series is the only (to my knowledge) analysis of Uber that assembled a complete picture of Uber’s profitability over time from the various fragmentary press reports about financial results. It argued that all the available factual evidence about Uber’s actual financial performance and competitive economics indicated that Uber’s business model could never produce sustainable profits unless it was able to exploit significant anti-competitive market power. None of the contrary claims made by Uber supporters have been backed by any objective economic data. Most media coverage totally ignores the abysmal economics and as a result can’t provide coherent explanations of Uber’s recent scandals and governance battles.
The material that had originally been presented across ten NC posts is now available in a single article published in the Transportation Law Journal and available at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2933177. The journal format allows for much more exhaustive documentation of evidence than internet posts, and makes it easier to demonstrate linkages between different aspects of the Uber story.
Newly released data affirm and strengthen my previous finding that Uber is hopelessly unprofitable
Since it is impossible to understand Uber’s behavior without first understanding its actual financial performance, I have updated the data that was presented in Section II-A of the TLJ article and in several parts of this NC series.
On December 1st, Eric Newcomer of Bloomberg published a new batch of Uber financial data that had been included in a prospectus related to SoftBank’s potential acquisition of Uber shares. The Uber historical P&L table below combines all available data and shows Uber full year results from 2012 to 2016; the estimated 2017 results are the actual 2nd/3rd quarter results reported by Newcomer multiplied by two. Uber has never challenged the accuracy of any of the P&L data that has appeared in the press.
The updated data demonstrates that Uber, in its eighth year of operations, continues to experience P&L losses that are staggering and still steadily growing, The data affirms that since there is no evidence of strong, ongoing profit improvement due to increasing efficiency or powerful scale/network economies, there is no basis for claiming that Uber can rapidly grow into profitability.
The data shows that while margin improvement occurred in 2015/16, it can be entirely explained by cuts in driver compensation unilaterally imposed by Uber, not by improved efficiency. New 2017 data strengthens the finding that the overall Uber business model (the combination of Uber and its nominally independent drivers) is not viable, and that all of Uber’s growth to date is due to billions in predatory subsidies. Those unilateral driver compensation cuts still left Uber billions short of breakeven but threatened the rapid growth its valuation was based on; reversing the cuts restored volume growth but pushed annual losses towards $5 billion.
Uber lost $2.5 billion in 2015, probably lost $4 billion in 2016, and is on track to lose $5 billion in 2017.
The top line on the table below shows is total passenger payments, which must be split between Uber corporate and its drivers. Driver gross earnings are substantially higher than actual take home pay, as gross earning must cover all the expenses drivers bear, including fuel, vehicle ownership, insurance and maintenance.
Most of the “profit” data released by Uber over time and discussed in the press is not true GAAP (generally accepted accounting principles) profit comparable to the net income numbers public companies publish but is EBIDTAR contribution. Companies have significant leeway as to how they calculate EBIDTAR (although it would exclude interest, taxes, depreciation, amortization) and the percentage of total costs excluded from EBIDTAR can vary significantly from quarter to quarter, given the impact of one-time expenses such as legal settlements and stock compensation. We only have true GAAP net profit results for 2014, 2015 and the 2nd/3rd quarters of 2017, but have EBIDTAR contribution numbers for all other periods.
Uber had GAAP net income of negative $2.6 billion in 2015, and a negative profit margin of 132%. This is consistent with the negative $2.0 billion loss and (143%) margin for the year ending September 2015 presented in part one of the NC Uber series over a year ago.
No GAAP profit results for 2016 have been disclosed, but actual losses likely exceed $4 billion given the EBIDTAR contribution of negative $3.2 billion. Uber’s GAAP losses for the 2nd and 3rd quarters of 2017 were over $2.5 billion, suggesting annual losses of roughly $5 billion.
While many Silicon Valley funded startups suffered large initial losses, none of them lost anything remotely close to $2.6 billion in their sixth year of operation and then doubled their losses to $5 billion in year eight. Reversing losses of this magnitude would require the greatest corporate financial turnaround in history.
No evidence of significant efficiency/scale gains; 2015 and 2016 margin improvements entirely explained by unilateral cuts in driver compensation, but losses soared when Uber had to reverse these cuts in 2017.
Total 2015 gross passenger payments were 200% higher than 2014, but Uber corporate revenue improved 300% because Uber cut the driver share of passenger revenue from 83% to 77%. This was an effective $500 million wealth transfer from drivers to Uber’s investors. These driver compensation cuts improved Uber’s EBIDTAR margin, but Uber’s P&L gains were wiped out by higher non-EBIDTAR expense. Thus the 300% Uber revenue growth did not result in any improvement in Uber profit margins.
In 2016, Uber unilaterally imposed much larger cuts in driver compensation, costing drivers an additional $3 billion. Prior to Uber’s market entry, the take home pay of big-city cab drivers in the US was in the $12-17/hour range, and these earnings were possible only if drivers worked 65-75 hours a week.
An independent study of the net earnings of Uber drivers (after accounting for the costs of the vehicles they had to provide) in Denver, Houston and Detroit in late 2015 (prior to Uber’s big 2016 cuts) found that driver earnings had fallen to the $10-13/hour range. Multiple recent news reports have documented how Uber drivers are increasing unable to support themselves from their reduced share of passenger payments.
A business model where profit improvement is hugely dependent on wage cuts is unsustainable, especially when take home wages fall to (or below) minimum wage levels. Uber’s primary focus has always been the rate of growth in gross passenger revenue, as this has been a major justification for its $68 billion valuation. This growth rate came under enormous pressure in 2017 given Uber efforts to raise fares, major increases in driver turnover as wages fell,  and the avalanche of adverse publicity it was facing.
Since mass driver defections would cause passenger volume growth to collapse completely , Uber was forced to reverse these cuts in 2017 and increased the driver share from 68% to 80%. This meant that Uber’s corporate revenue, which had grown over 300% in 2015 and over 200% in 2016 will probably only grow by about 15% in 2017.
If Uber had any ability to exploit the type of powerful efficiency and scale driven improvements that rapidly drove other tech companies towards sustainable profitability, one would see clear-cut evidence in this P&L table. There are undoubtedly a number of things Uber could do to reduce losses at the margin, but it is difficult to imagine it could suddenly find the $4-5 billion in profit improvement needed merely to reach breakeven.
The P&L data illustrates why Uber cannot go public
Under Travis Kalanick, Uber had no interest in an IPO because he fully understood that the full financial disclosures required—including historical cash flows, balance sheets and much greater operational P&L detail—would expose Uber’s abysmal economics, and destroy its PR narrative where powerful efficiencies would inevitably lead to success. Dara Khosrowshahi, under pressure from certain Board factions when he was first hired to replace Kalanick, promised an IPO by the end of 2019. This could be a disaster unless Uber somehow finds convincing evidence of profitable economics that it can put in the prospectus.
If Uber had accounting data that could demonstrate billions in efficiencies and a clear path to profits, they have ample incentive to share that data with reporters. Since they have not done that, it is reasonable to assume that evidence does not exist, and the additional data that would emerge during an IPO would actually strengthen the case that (in the absence of significant anti-competitive market power) Uber’s business model can never produce sustainable profits.
 Horan, Hubert, Will the Growth of Uber Increase Economic Welfare? 44 Transp. L.J., 33-105 (2017)
 That data, which had been leaked in dribs and drabs between August 2015 and April 2017, included the full years 2012-14, but only selected quarters in 2015 and 2016.
 Newcomer, Eric, Will Uber Ever Stop The Bleeding?, Bloomberg, 1 Dec 2017. Softbank’s tender offer is conditional on acquire a minimum of 14 percent of Uber shares at a price that would reflect a roughly 30% discount from Uber’s last valuation. All publically available data is strictly limited to high level corporate P&L numbers; no balance sheet, cash flow or regional/product level reports have been released.
. Data from the Softbank prospectus fills in the missing 2015 and 2016 data, and also provided results for the second and third quarter of 2017. There are some minor discrepancies between data released at different points in time; the P&L table uses the most recently released data. Limited first quarter of 2017 data had been released in May, but the Uber revenue number was wildly higher than previous or subsequent quarters. Bensinger, Greg, Uber Posts $708 Million Loss as Finance Head Leaves, Wall Street Journal, 1 Jun 2017
 Neither GAAP profit or EBIDTAR contribution includes expenses not directly related to current operations, such as long-term IT/market development costs, or expenses related to autonomous cars
 Given gross passenger payments of $20.0 billion in 2016, driver compensation was reduced by $3 billion due to Uber’s unilateral decision to cut the driver’s share from 83% (2014) to 68%
 See the TLJ article pp.46-49; the Denver/Detroit/Houston study is cited at note 37.
 One report cited the need for drivers to work marathon shifts focused on surge pricing periods. Masha Goncharova, Ride-Hailing Drivers are Slaves to the Surge, N.Y. Times (Jan. 12, 2017). Another report noted the increasing need for Uber drivers to actually sleep in their cars. Eric Newcomer & Olivia Zaleski, When Their Shifts End, Uber Drivers Set up Camp in Parking Lots across the U.S., Bloomberg News (Jan. 23, 2017). A third report confirmed the marathon shifts and sleeping in cars, and compared Uber drivers to “migrant workers.” See Carolyn Said, Long-Distance Uber, Lyft Drivers’ Crazy Commutes, Marathon Days, Big Paychecks, S.F. Chronicle (Feb. 18, 2017).
 Observed driver turnover would have been even higher, but most Uber drivers are locked into vehicle financing arrangements, and thus have no short-term ability to move to other jobs.