In this Article, we investigate the tax issues and challenges facing Uber and Lyft drivers by studying their online interactions in three internet discussion forums: Reddit.com, Uberpeople.net, and Intuit TurboTax AnswerXchange. Using descriptive statistics and content analysis, we examine (1) the substantive tax concerns facing forum participants, (2) how taxes affect their driving and profitability decisions, and (3) the degree of user sophistication, accuracy of legal advising, and other cultural features of the forums.
We find that while forum participants displayed generally accurate understandings of tax filing and income inclusion obligations, their approaches to expenses and deductions were less accurate and more varied in sophistication and willingness to comply with tax law. Forum participants also frequently discussed whether driving was profitable and exhibited a range of awareness concerning how taxes affected profitability. Finally, while the forums contained a surprising degree of sophisticated and accurate tax and legal advice, they also contained many examples of inaccurate or confusing information. It is thus uncertain whether forum readers were able to successfully distinguish between accurate and inaccurate advice dispensed in the forums.
Based on our findings, we make tentative recommendations for effective tax administration in the ridesharing and related sectors, including use of industry-specific guidance, clarification of how existing tax rules apply to ridesharing, and guidance on Form 1099-K interpretation. We analyze the implications of our findings regarding taxes and profitability for Uber’s business model and its potential regulation. Finally, we discuss the possible impacts of targeted tax compliance initiatives on internet communities.
- I. Introduction
- II. Understanding the Ridesharing Sector: Background and Research Questions
- III. The Study of Internet Forums: Research Methodology
- A. The Study of Internet Discussion Forums: Possibilities and Limitations
- B. Selection of Internet Forums for Study: Reddit, Uberpeople, and TurboTax
- C. Selection of Tax-Related Discussion Threads for Analysis
- D. Coding of Discussion Threads
- IV. The Tax Lives of Uber Drivers: Describing the Data
- V. The Tax Lives of Uber Drivers: Content Analysis of Data
- A. Tax Issues Confronting Ridesharing Drivers
- B. Role of Taxes in Decision Making
- C. Tax Sophistication and Tax Compliance Culture
- VI. Analysis and Recommendations: Compliance, Money, and Community in the Internet Forums
- VII. Conclusion
The past few years have seen the meteoric rise of Uber Technologies Inc. (Uber), a company that allows consumers to request rides from registered peer drivers using the Uber app. From its beginnings in San Francisco in 2009, Uber has grown to become one of the most highly valued startups in the world. According to one study, more than 460,000 drivers were “actively partnering” with Uber in the United States as of the end of 2015. As of June 18, 2016, Uber had provided two billion rides worldwide.
Uber, of course, is simply the most prominent example of a new model of production and consumption of goods and services often referred to as “sharing,” “collaborative consumption,” or “peer-to-peer consumption.” By employing internet and mobile phone platforms that reduce transaction costs, this model enables individuals to easily offer and enjoy goods and services—anything from homes and rides to bathrooms, Wi-Fi, and even kittens. As peer-to-peer consumption has proliferated, academics and policymakers have begun to tackle the legal and regulatory questions that it raises.
An issue of increasing importance is the impact of the peer-to-peer economy on the lives of those providing goods and services via these platforms. There is some evidence that this new economy has affected the nature of work and has resulted in the expansion of independent contractor jobs. Commentators refer to the rise of work through these platforms as the “1099 economy,” “independent contractor economy,” or “gig economy.” Academics, courts, regulators, and commentators are just now focusing on the legal issues confronting 1099 economy workers. For example, courts and tribunals are struggling with the question of whether ridesharing drivers should be classified as employees or independent contractors for purposes of labor and tax law.
In this Article, we contribute to the growing legal literature on 1099 or gig economy workers by studying the ways in which Uber and Lyft, Inc. (Lyft) drivers navigate the tax system. We investigate the tax issues and challenges faced by these drivers by examining the discussions and interactions that take place among drivers and other participants in three internet discussion forums: Uberpeople.net, the Uberdrivers subreddit on Reddit.com, and the Intuit TurboTax AnswerXchange Forum. Internet discussion forums can provide a timely picture of the tax and related issues that concern ridesharing drivers, particularly in an environment where drivers are learning how to navigate an emerging sector and where there may be a lag in obtaining tax return and other data. The anonymous nature of these forums may cause forum participants to be more forthright than they might be in surveys or in-person interviews. Substantive tax issues aside, online discussion forums are also interesting as a socio-legal phenomenon themselves. For example, they can serve as primary source evidence of how the process of seeking and giving tax advice functions on an internet discussion board.
We selected all tax-related discussions from each of these internet forums using a search methodology tailored to each forum. We then studied this self-assembled dataset using a combination of descriptive statistics and qualitative content analysis, focusing on three key research questions: (1) the main substantive tax issues and questions discussed by forum participants, (2) the ways taxes factor into forum participants’ analyses of profitability and their decisions to drive for ridesharing platforms, and (3) the dominant cultural attitudes displayed by forum participants, including their levels of sophistication, attitudes towards compliance and risk taking, and how they expressed disagreement and corrected each other’s errors.
We find that forum participants displayed generally accurate understandings regarding tax filing and income inclusion obligations. However, their approaches to expenses and deductions (particularly mileage calculations, expenses on the business-personal borderline, and documentation) were less accurate and more varied in sophistication and willingness to comply with tax law. Forum participants also engaged in frequent discussions about whether driving was ultimately profitable and exhibited a range of awareness of how taxes affected their bottom-line profitability. They used metrics developed by tax law (in particular, the IRS standard mileage rate) as rough proxies in estimating other expenses of driving. Tax burdens also played a role in shaping the broad online consensus that it was preferable to drive part time rather than full time. Finally, while the forums contained a surprising degree of sophisticated and accurate tax advice (including tax advice marketed by forum users claiming to be experts and professionals), they also contained many examples of inaccurate, confusing, or misleading statements about tax law. It is thus uncertain whether forum readers and participants are able to successfully distinguish between accurate and inaccurate advice dispensed on the discussion boards.
Based on our findings, we make three sets of recommendations and observations. First, we make suggestions for effective tax administration in the on-demand ridesharing and related sectors. The conversations we observed suggest that industry-specific IRS guidance, IRS guidance regarding which technologies and applications may be used to document mileage and other expense deductions, and IRS clarification with respect to Form 1099-K issuance and interpretation are fairly simple and uncontroversial steps that might help alleviate confusion and tax compliance burdens for drivers and other 1099 economy workers. The IRS recently took steps in this direction through the launch of an online “Sharing Economy Tax Center,” and more could be done in a similar vein. Second, we analyze the implications of our findings regarding how forum participants assess driver profitability (and the role played by taxes in making this determination) for Uber’s business model and for potential regulation of ridesharing. Finally, we discuss the role played by internet forums in facilitating learning about tax law and in disseminating legal information. We evaluate the promise and perils of intervening in or using internet forums to advance tax administration goals. For example, we consider how interventions in internet communities may have unintended impacts on the culture of such communities. While our insights and recommendations largely focus on the on-demand ridesharing sector, many of these observations may be extrapolated to other parts of the sharing economy as well.
This Article is the first in the legal literature to systematically mine internet discussion forums for detailed insights into how individuals learn about, interact with, and shape legal systems through online channels. It is also the first qualitative content analysis of online interactions among sharing economy workers as they grapple with questions of tax law in an emerging sector. By thickly describing how tax law plays out “in action” on internet discussion boards, our study opens a window into how ridesharing drivers and other 1099 workers interact and cope with complex tax and related regulatory regimes. This is particularly valuable because key decisions are being made about the legal and regulatory fate of such workers based on assumptions about worker preferences and their experiences with gig work, but these assumptions for the most part remain uninvestigated. By illuminating the ways in which these small-scale gig workers talk about, engage with, and confront the tax system, this Article contributes significantly to our understanding of the realities of 1099 economy work.
In Part II, we describe on-demand ridesharing and state our research questions. In Part III, we detail our research methodology, including how we chose the internet forums for analysis, how we selected discussion threads, and how we coded and analyzed the data. In Parts IV and V, we present and analyze our findings. In Part VI, we discuss the implications of our findings for tax administration, Uber’s business model, and regulation of ridesharing. We also examine the potential impacts of targeted tax compliance initiatives on internet communities and other design considerations.
In this Part, we describe the basic characteristics of the ridesharing sector and the major tax questions that impact this sector. We then describe our research questions.
The on-demand ridesharing sector involves the hailing of a vehicle for hire using a smartphone mobile app that relies on GPS technology to facilitate the pickup and ride. Uber is widely acknowledged as the market leader in this sector. The service is currently available in almost 200 U.S. cities and over 70 foreign countries. Uber’s basic business model involves partnering with local owners of licensed private car companies and with individuals driving their own personal vehicles. Uber does not own the cars, and the drivers themselves decide whether and when to open the mobile application and accept ride requests from customers. Drivers are responsible for their own expenses including gas, equipment maintenance, and repairs.
Uber does, however, set fares, based in part on the level of service provided. UberX is Uber’s best-known service and allows drivers to use their personal vehicles to offer rides to customers at fares often significantly lower than taxi fares for comparable trips. To apply to become an UberX driver all that is required is a driver’s license, a car that satisfies Uber’s requirements, proper insurance, and clearing a DMV and background check. Uber also offers other ride services in certain markets: UberBlack is a traditional “Black Car” service that resembles limousine services. In many U.S. cities, Uber riders also have the option of UberLUX, a luxury car service; UberSUV, a full-sized luxury SUV; UberTAXI, a licensed taxicab; UberXL, a non-luxury SUV; and UberPool, a reduced-fare pooled ride service.
A distinctive characteristic of Uber’s fare structure is its use of variable pricing levels, depending on demand. Under such dynamic or “surge” pricing, changes in fare price are determined algorithmically when wait times increase and unfulfilled requests start to rise. These fare increases may occur because of a demand surge during high traffic times or because of special conditions such as a holiday or inclement weather. At the time of our study, Uber told customers that its fares included an automatic 20% tip and that there is no need for an additional tip; but it now states that the bill does not include a tip and that tipping is voluntary.
In exchange for creating and providing the app-based marketplace for rides, Uber takes a portion of the gross fares (typically 20%, though this varies by city) generated by drivers and also deducts a “safe rides” fee. In a recently settled case, some Uber customers alleged that Uber also withheld part of the 20% tip that it had claimed was included in its fares, rather than paying all tips over to drivers. During our study’s timeframe, drivers could rent an Uber-ready smartphone from Uber but could also use their own phone and download the Uber smart phone application.
Lyft is Uber’s foremost competitor in the ridesharing market. Uber and Lyft are similar services and have nearly identical business models. Like Uber, Lyft connects passengers and drivers through Lyft’s smartphone application. Like Uber, Lyft offers a basic service (Lyft), a shared ride service (Lyft Line) and a six-passenger premium ride service (Lyft Plus). Like Uber, Lyft also elevates fares during periods when demand is high, and like Uber, Lyft provides a liability insurance policy for periods when a Lyft driver is ferrying a customer, as well as contingent liability for the period the Lyft application is on but the driver has not yet accepted a ride. There are some differences, however. For example, Lyft customers are prompted to pay within the Lyft application after the ride, and are able to tip the driver using the app, though a tip is not required. The Uber application currently does not allow for tipping. Lyft, like Uber, takes a 20% cut of the base fare; however, Lyft drivers keep 100% of all tips.
A third on-demand ridesharing service was Sidecar Technologies, Inc. (Sidecar), a ridesharing service that operated on a much smaller scale and was ultimately unable to compete. During 2015, Sidecar modified its business model to focus on delivery of goods rather than passenger rides but ultimately ceased operations in December 2015. Sidecar’s closure illustrates the changing landscape of the ridesharing sector, and leaves Uber and Lyft as the two main ridesharing services.
As a matter of formal tax law, the clear rule is that Uber drivers, like any other income earners, are required to include rideshare earnings in gross income but may deduct allowable expenses from their gross income. However, as we have argued elsewhere, tax compliance may raise certain challenges. For one thing, Uber classifies its drivers as independent contractors for tax and other purposes. As such, Uber does not withhold taxes on amounts paid to drivers and does not issue Forms W-2, but instead reports gross amounts earned by drivers to both the IRS and to drivers on Forms 1099-MISC and 1099-K. Such “information reporting” helps the IRS identify and corroborate income earned and reported by taxpayers on their tax returns, and studies suggest that in sectors where information reporting and withholding are difficult to impose (e.g., cash businesses), tax compliance declines.
Another wrinkle that may generate tax compliance consequences is that Uber and Lyft currently take different Form 1099-K reporting positions due to a perceived ambiguity in the law. Both platforms report gross driving payments earned by drivers (which are the bulk of driver payments) on Forms 1099-K but take different positions on reporting thresholds. The relevant statute requires that banks and other “merchant acquiring entities” report all payments made to payees, no matter how small, while “third party settlement organizations” must only report payments if they exceed $20,000 and if there are more than 200 transactions with the participating payee. Against this backdrop, Lyft has taken the position that it is a “third party settlement organization” only required to report payments above the $200,000/200 transactions threshold. Uber took this position until early 2015, but then began issuing Form 1099-K to drivers for all driving payments, no matter how small. The impact of Uber’s changed position and the resulting divergence between Uber and Lyft in their Form 1099-K information reporting can be observed in discussions on the internet boards.
Given the content of formal law, what tax issues do Uber and Lyft drivers actually discuss and debate on the internet? In this Article, we undertake this inquiry by examining three main questions:
1. Major Tax Concerns of Forum Participants
First, we investigate the major substantive tax questions actually discussed by forum participants. We identify the tax issues that are most relevant to forum participants and examine the ways in which they manage gaps, ambiguities, and compliance challenges that may arise. We also study the extent to which forum participants consult tax advisors, use tax preparation software, or employ other methods and resources in meeting their tax obligations.
2. Effect of Taxes in Decision-Making
Second, we investigate how forum participants make decisions regarding whether to drive at all and whether to drive part or full time, and we look at how taxes factor into those decisions. This second research question arose somewhat organically, because the data we collected contained many discussions of profitability and the role taxes play in its calculation. In addition to examining how forum participants talk about profitability overall, we specifically investigate when they begin to consider tax rates and administrative burdens (i.e., at outset, when deciding whether to drive, or later), whether they change their behaviors once tax and tax compliance burdens become salient, and whether their estimations and understandings of tax and compliance burdens are accurate.
3. Tax Sophistication, Tax Conduct, and Tax Compliance Culture
Finally, we analyze the dominant cultural characteristics of the discussion forums, examining features such as forum participants’ levels of sophistication in understanding tax law, their attitudes towards tax compliance and risk taking, the accuracy of tax information and advice exchanged in the forums, and the ways in which forum participants express disagreement and correct each other’s errors. We also describe the activities of non-driver forum posters on the discussion boards.
In designing our study, we made a series of methodological choices. In this Part, we discuss the key features of our research methodology: (1) the decision to study internet forums, (2) selection of the forums and discussion threads for analysis, and (3) the quantitative and qualitative analytical methods employed.
The use of discussion boards, chat rooms, blogs, and other internet forums to create community and exchange information is a well-established feature of contemporary life, and academics have increasingly realized that these internet forums are rich data sources that can be studied and analyzed. A growing body of literature in fields such as sociology and public health examines how posters in internet communities interact with each other and how information is conveyed through these online channels. However, detailed content analysis of how legal advice and information are transmitted through internet discussion forums has not, to our knowledge, been pursued. More specifically, in-depth analysis of how online communities of 1099 workers confront the tax system has never been performed.
Analysis of internet discussion forums is an appropriate method for studying ridesharing drivers and other sharing economy workers, particularly where other sources of data may not be immediately available. First, ridesharing drivers and other sharing economy workers make intensive use of internet and mobile technologies and are generally comfortable with online interaction and communication. Second, ridesharing drivers operate in comparative physical isolation and may have little in-person contact with other drivers. Thus, drivers have turned to the internet, converging in a number of online forums to share ideas and information. One news article accurately described Uberpeople.net, one of the forums we studied, as a “public water cooler” for atomized Uber drivers.
Third, the format and layout of internet discussion forums makes them an effective way to share technical information such as tax advice, original sources (e.g., links to other websites), spreadsheets, and legal material. Studying online forums therefore lets us observe the types of detailed technical information that are actually shared, stored, and possibly relied upon by forum participants and observers. Fourth, studying the online forums allows a real-time examination, based on a written transcript consisting of questions and comments, of what ridesharing drivers are contemplating and discussing. This has a clear benefit over methods such as surveys and interviews, which rely on driver recollections of past events. Fifth, the relative anonymity of online forums may encourage honesty regarding tax conduct and compliance behaviors. Finally, thick description of online interactions can add nuance and texture to our understanding of the research subject that other methodologies—such as quantitative data analysis—may fail to capture.
Yet, study of online discussion forums, while it might yield interesting insights, is subject to a number of limitations. Most obviously, only a subset of ridesharing drivers post on the discussion boards. We have limited information about how forum participants skew in terms of demographics, attitudes towards driving, or attitudes towards tax compliance. Thus, it is important not to take the views expressed on the forums as representative of the views of all drivers. Second, it is also difficult to know the impact of the discussion boards beyond the community of active forum participants. All of the boards we studied can be accessed and read by the public without logging in, so there are likely a significant number of “lurkers.” We did not obtain information about the number and frequency of lurkers or about site traffic, so the impact of information and opinions shared on the boards on such lurkers is unknown. Third, there may be a gap between what forum participants say anonymously on the internet and what they actually do in terms of tax compliance and other behaviors. We have no way to quantify that gap, if it exists. Thus, one must be careful in drawing conclusions about the actual behaviors of rideshare drivers based on what is said on the boards.
There are also necessary limitations in terms of study design and methodology. As with any qualitative study, certain aspects of how we categorized the discussion threads and coded the data might be subject to debate in close cases. In addition, our study is bounded by the dates on which we conducted the search, by our search terms, and by the search engines of the respective websites.
Notwithstanding these limitations, study of online discussion forums can yield valuable information and data, particularly when supplemented by other research methods.
We analyzed three internet discussion forums: the Uberdrivers subreddit on Reddit.com, the Uber drivers’ forum Uberpeople.net, and the Intuit TurboTax AnswerXchange Forum. The Uberdrivers subreddit is an important internet venue where drivers interact, with over 5,000 subscribed readers, while Uberpeople is a leading online forum for ridesharing drivers. While Reddit and Uberpeople are general interest sites, discussions on TurboTax pertain only to tax. We found, somewhat surprisingly, that the most substantive and sophisticated tax discussions were contained on Reddit and Uberpeople, not on TurboTax.
All three websites are in discussion board format, which means that individuals can post content in the forum by logging in under a username and writing an entry. Posted content often takes the form of questions or commentary. Other users, who may be members of the public, forum employees, or moderators, may respond to or comment on posted content by logging in under a username, thereby generating a discussion. In this Article, we refer to each discussion, which consists of the original post and subsequent comments and responses, as a “discussion thread.” We refer to the initial poster for each discussion thread as the “thread originator” or “original poster” and to subsequent posters on a thread as “commenters.” In all three cases, the websites were viewable by the general public without a username or a subscription; an account was required only to post content.
Reddit, which describes itself as “the front page of the internet,” is a widely used social networking and news website in which community members can post content by registering under a username and password. Reddit can be broadly characterized as a “bulletin board” upon which users post text-based entries, links, videos, photos, and other content. Content on Reddit is organized by communities of affinity or interest, known as “subreddits,” and subreddits range widely in tone, content, and topic. We observed a number of subreddits relating to the sharing economy. The Uberdrivers subreddit was by far the most significant and active subreddit for ridesharing drivers.
Reddit users may post content, comment on submitted content, and comment on those comments, thus generating a conversation within each discussion thread. Reddit operates a voting system, by which users (redditors) can vote up or down (“upvote” or “downvote”) on submitted content and comments. These upvotes and downvotes affect the ranking of items in each subreddit. During the study’s time frame, the content within a subreddit could be sorted based on the following categories: “relevance,” “new,” “hot,” “top,” and “comments.”
A Pew research survey conducted in April and May 2013 found that 6% of internet users surveyed use Reddit, and that men were twice as likely to be Reddit users (8%) as women (4%). Young people were more likely to use Reddit than older people: 11% of internet users ages 18-29 reported using Reddit while 7% of internet users ages 30-49 reported Reddit use. Younger men were significantly more likely to be Reddit users than older men. Urban and suburban dwellers were more likely to use Reddit than those living in rural areas.
By comparison, a 2016 study of Uber drivers commissioned by Uber concluded that 19.1 percent of drivers are under age 30, and only 21.8 percent are age 50 or older. In contrast, taxi drivers and chauffeurs are substantially older. Around 14 percent of Uber drivers are women. Although there is not a direct overlap between the demographic profile of Reddit users and Uber drivers, both seem to be predominantly male and young. Combined with the observation that Uber drivers are likely to be technologically savvy enough to use internet forums, it is reasonable to conclude that Reddit might provide useful, though by no means perfect, data on ridesharing drivers.
UberPeople.net is an internet discussion forum that identifies itself as “an independent community of rideshare drivers.” It is perhaps the leading internet discussion forum for ridesharing drivers and has been in operation since April 2014. The volume of discussion and the scope of ridesharing topics covered are extensive. The site’s ownership is unclear and has been the subject of some debate. Like Reddit, Uberpeople consists of discussion threads generated by a thread originator and commented upon by subsequent posters. Because of the forum layout, however, the Uberpeople discussion threads seem to have a slightly less conversational style than Reddit. It is generally accepted that some forum participants are self-identified Uber employees, and it is assumed that others may be unidentified Uber employees.
Intuit TurboTax AnswerXchange is an online question and answer forum created by Intuit, a leading tax preparation software developer that offers the TurboTax software. Users who have created an account can log into the forum and post questions about how to prepare their taxes. TurboTax employees, representatives, and other users provide publicly available responses to these questions. In contrast to Uberpeople and Reddit, all of the discussion threads on TurboTax concerned taxation, so we did not observe non-tax conversations morphing into discussions of tax issues (and vice versa).
Because of differences among the three forums, we tailored the discussion thread selection process for each forum to ensure that all relevant threads were identified. Both Uberpeople and Reddit contain a large number of non-tax-related discussions, so our goal was to isolate the tax-specific threads. The discussions on TurboTax all concerned taxation, so the goal there was to identify the threads relevant to ridesharing. Across all three forums, our objective was to create a database of tax conversations that could then be analyzed. We sought to explore the kinds of tax issues significant to posters, to understand how taxes fit into posters’ assessments of profitability, and to generate a picture of forum culture and sophistication and of how legal knowledge was transmitted on the discussion boards. We did not seek to make comparative claims about the dominance of tax discussions as compared to other types of discussions (for example, discussions about insurance). Therefore, our method of thread selection does not create a selection bias.
We searched for and downloaded tax-related threads in the Uberdrivers subreddit using the following words and phrases: taxes, tax, 1099-K, 1099K, 1099-misc, 1099misc, 1099, deduct AND expense, deduct* AND expense*, deduction*, taxation, deduct* AND mile*, deduct* AND cost*, mileage, IRS, “independent contractor,” and “independent contractors.” Our search methodology yielded 260 discussion threads.
We searched for tax-related threads on Uberpeople by combining a search term method and an individualized review of threads. The most pertinent section of the website was the “Forums” section, which contains the bulk of the discussion threads. “Forums” was itself divided into five categories: Community, Garage, Information, Options, and Geographical. During the study’s time period, these headings were further subdivided into subheadings.
We applied the following search procedure:
- The “Garage” category in the “Forums” section contained a subcategory “Pay” which itself contained two subcategories: “Taxes” and “Tips.” We included all the discussion threads under “Taxes” in our study.
- We performed a global search for the word “taxes” across the entire “Forums” section and included all hits in our study.
- We reviewed individually each remaining thread title in the “Forums” section and included all threads relevant to taxation in our study.
Our search methodology yielded 261 discussion threads.
We performed searches in the Intuit TurboTax Live Community Forum, using the terms “Uber” and “Lyft.” The search for “Uber” generated 85 threads, and the search for “Lyft” generated 22 threads, for a total of 107 threads.
We then subjected the selected discussion threads from the Reddit, Uberpeople, and TurboTax forums (n=260, n=261, and n=107, respectively) to quantitative analysis using descriptive statistics as well as qualitative content analysis. We sorted the threads in each forum into broad topical categories to get a general sense of the dominant topics of interest in the forums. We then analyzed the discussion threads under a system of 71 substantive codes and a number of other structural and in vivo codes using the data analysis software Atlas.TI. The full list of codes is in the Appendix. Coding the data helped us identify the most frequently discussed tax issues on the boards.
Substantive Codes. Our 71 substantive codes can be grouped into five broad families (total number of codes in each family in parentheses): Substantive Tax Issues (30), Tax Policy Issues (6), Attitudes Towards Tax Compliance (7), Tax-Related Strategies and Rationales (14), and Feelings and Emotions about Taxes (14).
Structural Codes. To obtain a more nuanced understanding of forum dynamics and composition, we employed a number of codes (between 7 and 11 depending on which forum) designed to illuminate the structure and layout of the forum. Thus, for example, we coded for thread-originating posts, posts by tax professionals and other marketing posts, possible “plants” by Uber and Lyft or by taxicab competitors, and possible web bots. The structural codes enabled us to tease out various features of the forums, such as number of questions asked, most commonly asked questions, and most frequent posters.
Thread Originators and Respondents. We coded the names of all forum participants. This allowed us to understand the number, frequency, and composition of forum users.
Other in Vivo Codes. Finally, where there were distinctive comments or interactions, we employed in vivo coding to capture precise language.
We first summarize our findings using descriptive statistics. We do so in order to give the reader a broad sense of the key topics discussed, the level of activity, and the composition of forum users in these internet discussion forums, before delving into our detailed content analysis.
The 260 discussion threads yielded by our search methodology were submitted between 27 November 2013 and 1 May 2015, with 109 of the 260 threads posted in 2015. Each thread (and each comment within a thread) is given a points score on Reddit, which is the number of “upvotes” minus the number of “downvotes” received. Each discussion thread also contains information about the thread’s “percentage upvoted,” which is the percentage of votes received that are upvotes.
Although all threads were selected based on some connection to tax (as described in Part III.C), the primary topics of the threads were not exclusively substantive tax issues. We therefore classified each thread into one of seven basic categories based on its primary focus. We then gathered data on (1) the date the thread was posted, (2) the number of comments on each thread, (3) the thread’s point score, (4) the percentage of upvotes received, and (4) the general composition of the forum users.
General Thread Topics. Based on our classification scheme, we found the following numerical distribution of topics: 
|Thread Topic||Number of Threads|
|Substantive Tax Issues||103|
|Uber’s Business Model||25|
|Interactions with Uber||15|
|Independent Contractor vs. Employee Issues||14|
This suggests that the dominant topic in the discussion forum threads pertained to tax-specific issues. This is unsurprising given our search terms. It also shows that a significant number of threads dealt with questions of whether rideshare driving was profitable and the role taxes played in such profitability determinations.
Number of Comments. The discussion threads contained an average of 13.45 comments per thread, with a median of 9.5 comments. There were a relatively small number of discussion threads that drew a large number of comments. Focusing on the 50 discussion threads that drew more than 20 comments per thread, the main topics of those threads were as follows:
|Topic||Number of Threads|
|Substantive Tax Issues||6|
|Uber’s Business Model||5|
|Independent Contractor vs. Employee Issues||3|
|Interactions with Uber||2|
This broadly suggests that even though there were more substantive tax threads than threads concerning profitability, the thread topics that drew the most commentary from redditors predominantly concerned whether rideshare driving is profitable, while substantive tax questions drew less interest.
Points Score and Percentage Upvotes. The scores of the discussion threads ranged from 118 to 0, with an average of 4.73 points and a median of 2 points. The percentage of upvotes received by each thread ranged from 100% to 18%, with an average of 73.75% upvoted and a median of 75%. We analyzed point scores and percentage of upvotes to figure out roughly which threads drew strong reactions and might have had a wider readership or impact.
Compositions of Forum Posters. Reddit forum posters were a combination of frequent, repeat posters and episodic or casual posters. The most frequent thread originator originated 11 threads. The handle USER176, which belongs to a marketer marketing the tryzen99.com website, was second with six threads originated. There was significant but imperfect overlap between the frequency of thread origination and the frequency with which each user commented or responded on the boards. For example, the most frequent thread originator in our dataset was the second most frequent thread commenter (81 occurrences). On the other hand, the most frequent thread commenter (86 occurrences) only originated one thread in our dataset. The handle USER176, who was the second most frequent thread originator, was the fifth most frequent respondent (63 occurrences). Thus, repeat commenters had a strong presence on the discussion board. In addition, USER176’s presence and frequent posts and comments supports our impression that the Reddit discussion board is a shared space between actual drivers and other parties, such as marketers, journalists, and suspected “plants” by ridesharing platforms and taxicab competitors.
Episodic or casual commenters were also present in large numbers. We recorded a total of 734 thread respondents (i.e., handles who actually responded to threads, rather than originating threads) in the forum. Out of these 734 respondents, 596 had posted comments on threads 5 times or fewer, 665 had posted comments on threads 10 times or fewer, and 710 had posted comments on threads 20 times or fewer. Thus, over 80% of respondents had posted comments 5 times or fewer. Only six respondents posted comments on threads more than 50 times, and only 24 respondents posed comments on threads more than 20 times. This suggests that the vast majority of commenters were casual, infrequent commenters, with a relatively small number of regulars.
The 261 discussion threads generated by our search methodology were initially submitted in a 12 month period, between 10 April 2014 and 24 April 2015. Of the 261 threads, 179 were posted in 2015 and 82 were posted in 2014. The Uberpeople forum includes a function that allows commenters to “like” a given post or comment, but this feature played a minor role compared to the Reddit upvote, downvote, and points system.
As with Reddit, we gathered data on (1) the date of the original post, (2) the number of comments on each thread, and (3) the general composition of forum users. We then classified each thread into the same seven basic categories used with respect to Reddit, but added an eighth “Miscellaneous” category because our search in Uberpeople.net yielded a few threads unrelated to any relevant category. (As with Reddit, the Uberpeople threads were selected for their relation to tax, but their dominant topics were not always substantive tax issues.)
General Thread Topics. Based on these eight categories, the 261 threads yielded by our search methodology were distributed as follows:
|Thread Topic||Number of Threads|
|Substantive Tax Issues||121|
|Uber’s Business Model||25|
|Interactions with Uber||19|
|Independent Contractor vs. Employee Issues||7|
Thus, as with Reddit, the dominant topic of discussion in the threads captured by our search criteria related to tax-specific issues, followed by discussions about profitability determinations. Discussions of Uber’s business model and driver interactions with the company were a small but notable component, but discussions of driving strategy featured slightly less prominently than on Reddit. There were slightly fewer marketing posts than on Reddit.
Number of Comments. The Uberpeople discussion threads contained an average of 23.64 comments and a median of 11 comments. Thus, the mean number of comments was significantly higher for Uberpeople than for Reddit, while the median was less dramatically different.
Looking at the 77 discussion threads that drew more than 20 comments per thread, the main topics of those threads (based on our eight category classification scheme) were as follows:
|Topic||Number of Threads|
|Substantive Tax Issues||18|
|Uber’s Business Model||9|
|Interactions with Uber||7|
|Independent Contractor vs. Employee Issues||5|
Thus, as was the case for Reddit, even though tax-specific threads were the largest category in the data (121 threads), they composed fewer of the frequently commented upon threads than profitability discussions.
Composition of Forum Users. Uberpeople forum users, like Reddit users, were a combination of repeat posters and casual or episodic ones. Of 261 discussion threads analyzed, the most frequent thread originator originated seven threads, and the second most frequent originated five. There was imperfect overlap between frequency of thread origination and frequency of comments: Out of the top 5 thread commenters in our dataset (number of comments: 196, 154, 154, 98, 97), one of these commenters was the most frequent thread originator (154 comments; originating 7 threads), but of the remaining four top commenters, three originated only one thread while the fourth did not originate any. The handle USER153 (owner of the tryzen99 website) originated three threads and made 67 comments.
While there were a fair number of repeat commenters (71 users made more than 20 comments in our dataset), the majority of forum participants were casual users. Out of 816 commenters, 581 commented 5 times or fewer, 673 commented 10 times or fewer, and 745 commented 20 times or fewer. Like Reddit, Uberpeople users appeared to consist of ridesharing drivers, marketers, journalists, tax professionals, taxi drivers, and others.
The search methodology outlined above yielded a total of 107 discussion threads. Each discussion thread consisted of a question asked by a thread originator, answers to the question, and comments. Some threads contained no answers and/or no comments. Unlike the Uberdrivers subreddit, there were almost no repeat posters. Out of a total of 99 thread originators, only eight of those thread originators posted more than one question, and all of those eight posted twice each. Ultimately, this forum proved less relevant to our analysis than Reddit and Uberpeople because of the limited number of answers and comments and because the dominant issues were return preparation logistics and software-specific questions.
Content analysis of the data reveals more detailed information about how forum posters navigate the tax system, make determinations about profitability and driving, and interact with each other on the discussion boards.
Forum participants discussed a number of substantive tax issues, with varying degrees of depth and accuracy. The frequency with which codes occurred provides a rough sense of the issues that were most discussed. In Reddit, some of the most commonly occurring substantive tax codes concerned business expense deductions: “Expenses and Deductibility” (437 occurrences), “Mileage” (388), “Documentation” (264), and “Method: Standard Mileage vs. Other” (248). Issues surrounding drivers’ status as 1099 workers also emerged as an important topic: “Employee or Independent Contractor” (182), “Form 1099: Content and Accuracy” (69), “Form 1099: Different Reporting Positions” (18), “Form 1099: Receipt” (80), “Form 1099: Residual” (52), “Employment Taxes” (41), “Estimated Taxes” (43), “Withholding” (57), and “Schedules” (39).
The frequency of codes was similar in Uberpeople. Codes related to business expenses included: “Mileage” (678), “Expenses and Deductibility” (649), “Documentation” (280), and “Method: Standard Mileage vs. Other” (271). Codes related to driver classification included: “Employee or Independent Contractor” (295), “Form 1099: Content and Accuracy” (142), and “Form 1099: Receipt” (126).
At a broad level, the incidence of codes suggests that when forum posters talk about tax they are generally focused on questions and problems pertaining to business deductions and to the implications of their status as small business operators or Form 1099 recipients. To be sure, there were some discussions regarding the need to include income and some instances of failure or unwillingness to report income. For example, there were a few cases where posters filed their tax returns for the 2014 tax year in January 2015, “forgot” to include ridesharing income, subsequently received Forms 1099-K from Uber, and asked how to amend their returns. However, these income inclusions questions were relatively few compared to questions about expense taking. Most forum participants seemed to understand the basic requirement that they pay taxes on rideshare income. The observation that most forum participants view rideshare income as reportable seems broadly consistent with findings from existing tax compliance literature that third party information reporting (such as reporting on Form 1099-K) can improve taxpayer reporting of income. Drivers’ receipt of Forms 1099-K from Uber, combined with their knowledge that the entire business is conducted via credit card, might explain the widespread acceptance of the obligation to include driving income on the tax return.
On the other hand, examination of the discussion threads reveals that forum posters struggled with how to handle their business expenses. Many posters were new to filing taxes as independent contractors. Accordingly, issues surrounding Schedule C filing and expense taking and documentation were often matters of first impression. Specifically, drivers focused on the following core concerns: (a) determination of business miles and deductibility of expenses, (b) the implications of the choice between the standard mileage rate and the actual expenses method, (c) documentation, (d) understanding their Forms 1099-K, and (e) tax preparation issues and logistics.
Most forum posters in the Uberpeople and Reddit forums seemed to understand that certain types of expenses could potentially be deducted, readily identifying gas, bottled water for customers, and tolls as deductible. However, posters displayed uncertainty regarding: (a) line drawing between business and personal outlays, and (b) the timing of deductions and whether expenditures had to be capitalized instead.
a. The Business-Personal Borderline: Mileage and Other Expenses
With respect to the business-personal borderline, a frequent issue was whether miles spent driving without a passenger in the car counted as miles driven for ridesharing. Some background information is in order: First, the tax law gives drivers a choice between taking the standard mileage deduction (i.e., an IRS-specified deduction equal to $0.56 per mile driven on business for 2014) or deducting actual expenses incurred in business driving. However, if there is any personal use of the vehicle, then the question of which miles “count” is material no matter which method the taxpayer chooses, and she would have to track and document miles driven for ridesharing as opposed to personal use. Second, the rules for determining deductibility of business mileage are complicated. Generally, only costs associated with miles driven for business are deductible. Miles driven from “home” to work and vice versa, referred to as “commuting” miles, are generally not treated as business miles. However, once “at work” costs attributable to miles driven for work, for example, traveling between business locations, would be deductible. As the discussion below illustrates, a number of interpretative issues arise in applying these rules to ridesharing.
Uninformed forum posters expressed surprise that they needed to keep track of mileage information at all and frustration at having to do so. Interestingly, however, not many drivers appeared completely uninformed. A few drivers, who seemed to sense that miles were important but had not given the matter much thought, anticipated relying on Uber-provided statements regarding miles driven. However, Uber’s statements only include miles driven with a passenger.
More sophisticated drivers realized that the Uber statement numbers were likely under-inclusive of legitimate business miles. In that case, drivers would need to make their own determinations of business miles. Many posters assumed that all miles driving around with the Uber app on would “count.” Some forum posters contended that if the driver turned on the Uber app upon leaving home, then the driver had effectively started working and could count the mileage as business related. Even more aggressively, at least one commenter suggested that a driver is “working” whenever the app is on, regardless of whether the driver was really planning to work. These opinions might not be correct, however, because the IRS might take the position that miles driven from home to the usual place of passenger pickup constitute nondeductible commuting miles.
The most sophisticated discussions of mileage did not presume that simply turning on the Uber app was sufficient to render the miles driven deductible. Instead, these discussions focused on whether there was a notable personal component to the miles driven. For example, some posters wondered whether miles driven from home to either the first passenger pickup or the general pickup zone were really business miles, or whether those constituted nondeductible commuting miles. But even in these discussions there was confusion among some drivers as to the potential distinction between commuting miles (e.g., driving from home to “work”) and “dead miles” (i.e., driving around between passengers without a passenger in the car). The latter are likely equivalent to miles driven between clients, which are deductible under current law, but this was not obvious to all forum participants.
In some cases, the forum debates took on a style familiar to legal reasoning–analogizing the current case to established, more familiar ones. One commenter offered a broader vision of the underlying tax principle at stake:
The question the IRS will want you to answer is whether those miles were driven primarily for business purposes. Practically speaking you can do whatever you want and just count on not being audited. But if you are audited and you’ve been fudging your mileage the IRS will not be amused.
If you want to do your taxes honestly, consider the following examples: 1. Driving to do an errand with the Uber app on. Your primary purpose is running on [sic] errand. This mileage should not be deducted. 2. Driving to a busy part of town with the intent to pick up Uber passengers (even if you get no fares). Deduct this.
Interestingly, the debate over which miles count never seemed fully resolved (regardless of the legal accuracy of any such resolution) on the discussion boards. Moreover, it also did not appear resolved among advisors and self-identified experts: among posters who had sought professional tax advice, there seemed to be variation in the tax advice given (or at least variation in how the drivers understood that advice). We note also that the handle USER176, a former CPA who was offering tax advice to drivers on the forums and through his website and startup, at one point took the position that commuting miles could be deducted,  though he appeared to back off from this position slightly in posts on Uberpeople.
Finally, forum discussions revealed that some drivers adopted a strategy of driving home to wait between passenger rides. Regardless of whether miles driven between home and the city center to begin or end the driving workday, or miles driven from home to the first passenger pickup, are deductible, the analysis might be more complicated for drivers who repeatedly head home within the period of their self-designated driving shift.
Another less significant topic of discussion was the deductibility of clothing and related costs. Several posters mentioned deducting clothing and dry-cleaning costs, particularly for UberBlack, a limousine-like service. However, other posters would usually reply to inform them that clothing was not deductible. The same dynamic occurred with respect to meals. For example, one driver inquired whether meals would be deductible if the taxpayer opted to use the standard mileage method (essentially a question of whether meal expenses were subsumed under the standard mileage safe harbor deduction or could be additionally deducted). The first reply stated:
Meals do not seem like they would qualify: According to the IRS, “travel expenses are the ordinary and necessary expenses of traveling away from home for your business, profession, or job.” However, for these expenses to be deductible, you must be away from your tax home (the town where your business is located) for a period longer than one day’s work, and you must be away overnight . . .
The process by which forum posters corrected each other’s errors and misunderstandings is discussed further below.
Notably, most of the discussion of mileage and other expense deductions was found in the Uberpeople and Reddit forums. With few exceptions, TurboTax contained a more limited discussion of expenses and less extensive debate regarding which miles counted. The questions posted were generally quite basic, and there were fewer nuanced questions about expenses and mileage. Some of the answers were more sophisticated, but many were vague or contained standard or form language that did not adequately tailor the advice to ridesharing. In one case, for example, in response to a question about what can be deducted, the answer provided offered a good overview of the law, but did not address the law’s specific application to ridesharing. In another, a “TurboTax TaxPro” simply stated that drivers could rely on the statement provided by Uber but noted that “your mileage may actually be more as you drove around waiting for a hit.” Elsewhere, that poster recommended a method of estimating miles driven whereby the driver “keep[s] track of your mileage for one month starting from when you go available to when you go offline.”
The fact that forum posters struggled with expenses and deductions—but were relatively compliant regarding income inclusions—may dovetail with the predictions of the economics literature on tax compliance. That literature suggests that even if Form 1099-K reporting prompts taxpayers to accurately state income on their tax returns, such compliance may be partially eroded by inaccurate or excessive expense taking by the same taxpayers. We discuss the possible implications in greater detail in Part VI.
b. Timing Issues
Mileage and other expenses aside, some forum participants also displayed a limited understanding of timing issues and cost recovery. Specifically, to the extent that the financial expenditure involved something other than a classic recurring outlay (e.g., gas, tolls, etc.), some comments suggested that posters did not appreciate the difference between current deductibility of the entire outlay and depreciation or capitalization over time. For example, some posters erroneously raised the possibility of deducting the purchase price of the vehicle. Overall, this was not a dominant problem, perhaps because most drivers were using a preexisting personal vehicle for ridesharing. However, in cases in which drivers were contemplating the purchase or lease of a car to use for ridesharing, misunderstandings regarding deductions versus capitalization of outlays may present a bigger problem.
Another common topic of discussion concerned the choice, provided by tax law, of the standard mileage method as opposed to the “actual expenses” method of recovering the costs of using a vehicle for business. As described above, the standard mileage method allows a set per mile deduction ($0.56/mile for 2014) for the costs of automobile operation while the actual expenses method requires tallying up allowable actual costs allocable to business use. If the driver takes the standard mileage deduction, certain non-automobile business costs may also be deductible if the requirements of the relevant statutes are satisfied.
Once again the questions and answers reflected a significant range in awareness and sophistication. Most posters on Reddit and Uberpeople appeared familiar with at least the basic idea of the standard mileage option. Those new to driving and the forum were quickly informed of the method and the need, regardless of method, to maintain mileage logs. In a few cases, the description of the method was sufficiently sophisticated so as to include a discussion of limitations on the driver’s subsequent ability to switch methods. Posters in both Reddit and Uberpeople generally agreed that drivers were better off using the standard mileage method as opposed to deducting actual expenses.
A more nuanced question concerned which deductions a driver could additionally deduct if they selected the standard mileage method. Forum posters generally agreed (correctly) that some other expenses could be deducted. There was even general agreement regarding the ability to deduct outlays such as tolls, fees to Uber, Uber phone rental, and water and mints for passengers. Beyond that, however, posters argued over whether a wide variety of additional expenditures could be deducted in addition to the standard mileage deduction, including car washes, driver meals, driver clothing, and car detailing. For example, in response to one poster’s question as to whether the decision to deduct car washes was appropriate, another observed: “I’m fairly sure it as [sic]. As I understand it anything not directly mileage related can be deducted in addition to the standard deduction. SO no oil change deductions because it’s mileage related but washes and water for example are OK.” In contrast, another poster observed: “I like to have an idea of how much I’m actually making. The 56 cents per mile is a reasonable estimate of my expenses in terms of fuel, vehicle depreciation, and wear-and-tear, washes, etc.” Still other posters appeared divided on whether car washes could be deducted.
Perhaps one of the most dominant refrains through the Reddit and Uberpeople forums concerned the need to keep track of and retain documentation for expenses taken, most importantly mileage. As noted above, most forum posters understood that mileage numbers were important in preparing their tax returns, and many realized that the mileage numbers provided by Uber and Lyft would understate permissible business miles and limit their tax deductions. As a result, posters devoted a substantial amount of time to discussing documentation, best practices, and technology that might facilitate good documentation. Although a few outliers indicated that they would “estimate” the needed numbers, most posters were concerned with determining the best way to contemporaneously document mileage and other expenses. This suggested a relatively compliant ethos in the forums. The risk of IRS audit was a palpable concern in forum discussions, though one that posters might have overestimated. Forum participants actively considered what would constitute adequate documentation if one were audited.
Forum participants liberally shared information and suggestions regarding the use of various apps and spreadsheets to help track, store, and document mileage and expenditures. The Reddit and Uberpeople datasets each included over 100 coded references to third party apps including Expensify, MileIQ, Metromile, Google Sheets, GasBuddy, Sherpashare, Dash, and Triplog. The discussions about using these apps reflected worries about audit risk. Thus, for example, thread participants discussed evidence that, upon audit, the IRS might prefer paper records for mileage documentation.
As noted, what is not discernable from the forum discussions is the profile of and effects of these discussions on the lurking reader who does not post. For example, it is uncertain whether such readers comply with documentation requirements at the level described by forum posters. We also cannot tell whether posters themselves actually comply in the manner they publicly describe. Nonetheless, at least the public conversation viewable by members of the ridesharing community (both posters and lurkers) devotes a notable amount of time and attention to issues of tax compliance and documentation.
Drivers also showed varying levels of understanding of how income was reported on the Forms 1099-K received from Uber in early 2015. As discussed above, this was the year in which Uber changed its reporting position on Form 1099-K to report all driving payments regardless of size. This change generated confusion across all three forums. A few posters were puzzled that they received Form 1099-K from Uber in 2015 at all, because this was a departure from Uber’s prior reporting position. Additionally, some posters had questions about the numbers reported on Form 1099-K. Some correctly understood that the Forms 1099-K listed gross amounts, including the safe rides fees, tolls, Uber fees, and split fare fees, and that such fees should be deducted on Schedule C, Line 10 of the tax return to generate the appropriate net number. Others were confused about why Form 1099-K included gross amounts and thought that the forms contained errors. In all three forums, a number of posts debated the accuracy of amounts reported on Form 1099-K and how to adjust for fees and commissions. Eventually, more knowledgeable forum participants (including marketers like USER176/USER153) intervened to explain correctly how to interpret the Forms 1099-K. However, the impact of such interventions is not certain. As we discuss in Part VI, these data raise the possibility, not previously explored in the information reporting literature, that some Form 1099-K recipients may actually have over-reported earnings by failing to understand that amounts earned from Uber were gross amounts.
Relatedly, some posters who drove for Lyft seemed unsure of how to report income, in cases where they did not receive Forms 1099 from Lyft. Others were also confused as to why they received Form 1099-K from Uber but not from Lyft. Some of these posters raised questions of how and whether to include income from driving for Lyft.
While not predominantly related to taxation, driver opinions regarding worker classification warranted attention due to their tax implications. As discussed, the threshold decision by the ridesharing platforms to classify workers as independent contractors rather than employees gave rise to the issuance of Forms 1099-K and 1099-MISC and to accompanying issues such as expense taking and Schedule C filing. Despite this causal connection between independent contractor classification and the receipt of Forms 1099 and related tax compliance and reporting issues, forum posters did not typically associate their opinions about worker classification with these tax compliance matters. There were one or two notable exceptions, which we discuss below.
Some forum participants were enthusiastic about the prospect of being classified as independent contractors rather than employees while others were not. But in both cases, tax reporting and compliance issues were not the dominant reasons for their expressed preferences; rather, their opinions were based on non-tax reasons. For example, on one discussion thread dealing primarily with the employee-independent contractor debate, out of 37 posts, we saw 13 comments by 7 different posters expressing a preference for independent contractor rather than employee classification. Reasons for preferring independent contractor classification ranged from no stated reason to concerns about being liable for back taxes, to preferences for “not having a boss,” to flexibility and independence and not having to report to a superior. Other than the thread originator, there was only one poster who expressed enthusiasm for classification as an employee. The thread originator cited ability to make claims for back taxes and other benefits (including overtime and unemployment benefits) as upsides of employee classification.
These data suggest that there is heterogeneity in classification preference but also a range of underlying rationales for those classification preferences. Notably, however, conversations about the potential tax compliance and reporting implications of independent contractor vs. employee classification were generally absent. The meaning of the posters’ references to back taxes was unclear, but such concerns about back taxes were raised by posters on both sides of the debate. We speculate that some forum posters were worried that authorities would pursue them for self-employment taxes that had not been paid. Other than the back taxes reference, however, forum posters did not causally associate worker classification with the tax compliance and reporting issues they were facing, but rather based their preference on other non-tax factors.
Forum participants discussed and debated how to prepare and file taxes and what modalities and instruments to use in doing so. Several forum posters consulted, thought about consulting, or encouraged others to consult a tax preparer or advisor. We counted 115 coded references to tax preparers or advisors on Reddit and 196 on Uberpeople. One poster hilariously summarized the types of concerns that might have motivated some drivers to seek professional tax help:
. . . Get yourself a good accountant, that’s what I did. Cost me $300 a year but saves my butt. Remember, the IRS is the only agency where you are GUILTY and must prove yourself INNOCENT! No other government agency has that power, it’s the one agency that I won’t screw with.
NSA, CIA, FBI, ATF. go jump off a cliff. IRS, yes sir, how can I help you sir, yes sir I’ll do that sir, you bet sir . . . 
Other posters expressed frustration regarding the costs of such tax preparation help or were otherwise sensitive to the costs of hiring an adviser or preparer. Among forum participants who used a tax preparer or advisor, there was some appreciation of variation in the quality and skill of such preparers or advisors.
Other forum posters stated that they prepared their own taxes using tax preparation software. We counted 44 references to tax preparation software on Reddit and 104 on Uberpeople. One issue that was evident from TurboTax was the question of what version of TurboTax software was appropriate. 14 out of 107 discussion threads in TurboTax dealt with the question of which version of the software to use and whether rideshare drivers needed to upgrade to the TurboTax Home and Business version to prepare their tax return. Some posters appeared surprised that they needed to upgrade from TurboTax Deluxe to Home and Business to file Schedule C. Interestingly, a factor in their surprise or confusion seemed to be the “soft” sense held by some forum posters that they were not “running a business” and so should not be using the Home and Business version. The discussions about software versions and upgrades tracked the well-known TurboTax product controversy of January 2015, in which TurboTax customers were displeased at having to pay to upgrade to TurboTax Home and Business desktop software to file certain Schedules (including Schedule C). TurboTax reportedly backtracked and offered Deluxe customers who had to upgrade a $25 rebate. However, the fact of the rebate was not reflected in the TurboTax forum.
Another important theme was how taxes factored into forum posters’ assessments regarding profitability of rideshare driving and decisions whether or not to drive. Taxes factored into driving decisions and profitability in two ways: (a) decisions whether to drive at all, and (b) decisions about possible driving strategies.
As discussed in Part IV, profitability was among the most frequent discussion topics in the threads we examined, and threads discussing profitability drew some of the largest numbers of comments. Some participants posed the profitability question prior to deciding to drive. Others, having driven, sought advice on how to assess their own profitability. Still others shared detailed calculations of their profitability computations. Forum posters considered several different elements in their computations of profitability, including expenses, fees, tolls, taxes, and wear and tear on the vehicle. The debates regarding how to evaluate profitability were never fully resolved, in part because different participants appeared to use different assumptions, facts, and terminology.
Taxes fit into these conversations in three ways: (1) as taxes per se, (2) as a proxy for the costs of rideshare driving, and (3) as an indirect framing device that affected whether drivers viewed profitability on a per mile or a per hour basis.
a. Taxes as Taxes
Although most posters understood at least at a broad level that taxes would reduce their year-end take-home cash, a few did not appreciate the full potential impact of taxes on profitability. Some, for example, did not fully understand their obligation to “withhold” taxes for themselves (both income tax and employment taxes) and the possible need to file estimated taxes on a quarterly basis during the year. One poster summed up the point succinctly: “I wonder how many people will stop driving for Uber once they get their 1099 and do their taxes? I’m guessing quite a few will once they realize how much money they’re actually making. Thoughts?” Other forum posters seemed to appreciate the need to set aside taxes. Still others opined (sarcastically) that given the low Uber rates and expense deductions, they would not have to worry about owing taxes. Thus, there was a range in forum poster awareness of the impact of tax burdens.
The fact that some posters did not appreciate the full impact of taxes on their profitability may reflect an instance of what the literature on tax salience calls taxpayer “spotlighting” on gross prices. That literature broadly suggests that how a tax is presented or perceived will affect taxpayer behavior in ways different from what one might expect in a perfectly rational world. The spotlighting hypothesis suggests that taxpayers may underestimate the full tax burden and simply focus on the immediate price schedule in front of them. Spotlighting can arise when the imposition of the tax is separate from the market decision, for example, where the decision to drive occurs before receiving the tax bill. Scholars hypothesize that spotlighting might also affect job choices and work decisions. The question of how taxpayers respond to perceived tax burdens may be particularly important for the sharing economy because of the presence of new entrants and because labor supply in this sector may be relatively elastic.
Relatedly, we observed estimating/guesstimating of the overall federal and state tax burdens, with little analysis of marginal rates. Several forum participants seemed to pick a feasible sounding average number in “guesstimating” how much to put away over the course of the year to pay the eventual federal and state tax bills. When used in this manner, an estimated average tax rate can be a useful financial planning heuristic. However, to the extent drivers are relying on calculations of after-tax returns in making labor supply decisions, their decision-making process may be broadly consistent with another prediction of the salience literature, the “ironing” hypothesis. Ironing predicts that when faced with a progressive income tax or complex rate schedules, taxpayers may focus on average rather than marginal tax rates when assessing the profitability of an economic activity (such as consumption or supplying labor). It is difficult to distinguish between rough financial planning uses of average rates and actual ironing behavior based on the internet board discussions. However, it is worth noting that most posters did not demonstrate awareness of the concept of a marginal tax rate at all and did not appear to rely on marginal rates in making driving decisions.
In addition to the role of actual tax burdens in determining profitability, tax compliance costs were not lost on forum participants. One participant posted the following detailed analysis regarding compliance costs in the ridesharing sector:
There are costs associated with doing all this tax compliance correctly andd [sic] they are a deductible expense, but an expense and tax filing obligation that does not scale well for a one car fleet. Uber has turned the game upsidedown. When I owned and managed businesses, I had an accountant on retainer and a fulltime book keeper. I had assets, tow trucks and rental cars at risk. Uber has put its independent contractor drivers in the situations of assets at risk and tax compliance obligations and expenses while it keeps the reward potential that went hand in hand with assets at risk and compliance expenses, mostly in the lap of its drivers, without considering the disincentives for drivers to invest and put the assets invested in at risk. Uber sets the pricing. The driver cannot work harder to increase income because his costs mostly rise proportionately with extra miles driven to gross higher income. Drivers are attracted by a near instant direct deposit and avoid considering all of the risks and costs vs. the reward potential. It is an unsustainable business model, great for Uber and Uberx riders as long as the service is available anywhere near current pricing.
Thus, at least a few drivers thought about the costs of tax preparation and compliance for independent contractors as a factor in determining whether rideshare driving was profitable.
One final point bears noting: as discussed above, most posters did not directly connect worker classification issues with tax filing and compliance obligations. But some did indirectly raise a connection between the status of drivers as employees or independent contractors and the potential impact on the ridesharing business model and its ultimately profitability (with taxes being a component of such profitability). Mixed feelings and opinions about worker classification were evident in various discussion threads and often seemed tied to uncertainty about the actual economic implications of one classification over another. For example, some posters questioned whether litigation would force Uber to treat drivers as “true” independent contractors by shifting more costs to drivers, while others maintained that even a finding that drivers were employees would have an indeterminate economic effect on actual driver costs. Some posters expressed doubt that employee status would improve their situation because the costs would ultimately be borne by drivers anyway. Thus, even though the posters’ worker classification comments did not pertain directly to tax law, the indirect economic implications (which encompasses tax burdens) did not escape forum posters.
b. Tax Numbers as Proxies
Tax numbers also played a role as proxies in how forum posters estimated other expenses of driving. A major source of difficulty for forum posters was how to calculate and evaluate the costs of using personal property for business, including costs associated with vehicle wear and tear. In this context, many drivers used the tax system’s standard mileage allowance ($.56/mile for 2014) as a proxy for calculating economic costs of driving the vehicle, especially in accounting for the impact of depreciation. Several posters appeared not to understand that the standard mileage allowance was a tax number and not necessarily a reflection of economic reality. Others criticized the use of the standard mileage allowance as irrelevant for purposes of determining economic profitability, arguing that it actually costs nobody $.56/mile to operate their vehicle.
More generally, forum participants displayed confusion and disagreement regarding the relationship between tax losses and economic losses. Some posters expressed satisfaction at incurring a tax loss, not seeming to realize that the tax loss might reflect a true economic loss. Conversely, others failed to appreciate that tax losses might be generated through aggressive tax reporting by drivers experiencing actual economic profits. Consider the following contrasting responses to a forum poster who outlined how, after taking the standard mileage deduction and other deductions, there was no taxable income from Uber driving:
USER124: “The milage [sic] deduction to me was a huge item. I used Turbo Tax and Quickbooks to enter all items and as far as I’m concerned I made money just by not having to pay a large tax bill like I did last year before I started driving Uber.”
USER56: “If this goes through, for you and for others, this will represent one of the greatest scams on American tax payers, in years.”
USER94: “Only Uber drivers are dumb enough to be excited about pulling a loss on their operations.”
This exchange illustrates how the first two posters regard tax losses as an economic positive for Uber drivers, believing that those tax losses are disguising economic gains, but the third presumes that those tax losses reflect actual economic losses. Of course, there is no way of knowing who is right without additional information.
c. Taxes as a Framing Device for Profitability Computations
Tax concepts also influenced whether forum participants measured and discussed profitability on a per-mile or a per-hour basis. Forum participants frequently made profitability calculations on a per mile basis. The per-mile framing was likely prompted by how two key costs of driving are measured: (1) tax law’s standard mileage deduction, which is frequently used as a proxy for assessing profitability, and (2) gas and vehicle mileage, which is computed on a MPG basis. However, sometimes drivers also described their profitability in dollars per hour, reflecting their frequent comparisons of the decision to drive with alternative low wage employment, e.g., Walmart, which pays on a per-hour basis. There were fewer references to per-week or per-month computations.
a. The Role of Taxes in Deciding “How” to Drive
Beyond the binary decision of whether to drive at all, drivers face a secondary decision of “how” to drive, that is, whether to drive on a part-time or full-time basis. Here, again, taxes played a role in decision making both in their own right and as proxies for other important measurements. There seemed to be a general consensus that it is more likely to make sense to drive for Uber part time than full time. This consensus was supported by a number of non-tax factors including: (1) driving part time enables the driver to select the most profitable times of day to drive whereas full-time drivers need to drive at suboptimal times and are more likely to experience burnout, (2) forum participants perceive vehicle wear and tear as less significant and feel more justified in ignoring these costs when driving part time, and (3) the instability of Uber’s pricing/rate structure and the risk of relying on Uber to make a living. Tax factors also played a role. Forum participants judged obligations for quarterly filing and payment of estimated taxes to be less onerous for part-time drivers with other full-time employment because such drivers could avoid estimated tax obligations by simply increasing wage withholding from their W-2 jobs. Forum participants also seemed to appreciate, implicitly, the value of untaxed benefits (e.g., paid vacation) received from alternative wage employment.
Although part-time driving was viewed as preferable to full-time driving, some forum participants noted that even part-time driving might not be profitable. One poster wrote:
. . . I see people as being in 2 categories either: a) You are trying to make Uber a full time job b) You have a job and use Uber to supplement your income . . . If you are in category (b), you will probably make around $12 an hour if you live in a decent city, or more if you time your hours so you only work during the most profitable times (full timers don’t have this luxury obviously). But if you live anywhere else you will be making below minimum wage. At that point it would only make sense to continue driving if you literally can’t make ends meet and you can’t find another job to do on the weekends.
Uber knows this and exploits the fact that its parttime drivers are either desperate for money or say “it’s for beer money” and don’t bother to do the calculations to realize they are making below minimum wage. And if that’s the case, you would be better off working at McDonalds. At least that job is here to stay, you know exactly how much you are making, and you aren’t putting miles on your car.
The sentiment here is that part-time drivers may not have the incentive to pay attention to taxes and other expenses in determining whether driving is profitable.
b. The Labor-Leisure Trade-Off in Part-Time Driving
Discussions among forum posters regarding their decisions to drive part time reveal interesting insights into the labor-leisure trade-off, a frequent subject of economic analysis. Forum posters generally did not explicitly discuss leisure as something that they traded off against driving. However, they seemed to weigh leisure time implicitly, as reflected by: (1) discussions about the flexibility of driving for Uber, (2) descriptions of driving as having both business and pleasure components for a subset of forum participants, and (3) descriptions of the incentive to drive as the opportunity to earn some money for “extras” or “beer money.” These descriptions are suggestive of driving being a choice, to be weighed against other alternatives, such as “free time” on the weekend. As discussed above, it is possible that forum posters made these choices while mismeasuring the true costs of their earnings (through the failure to fully account for expenses, taxes, and wear and tear).
These observations about forum posters’ profitability assessments, driving decisions, and labor-leisure trade-offs help illuminate how cognitive limitations and market considerations (including tax considerations) combine to shape their labor supply decisions. For example, to the extent that mismeasurement of driving costs affects driver decisions, it is unclear whether having a full year’s experience driving is enough to completely mitigate the mismeasurement. Some costs might become visible during the tax return preparation process and allow drivers to revise their profitability assessments or implement new strategies (e.g., setting aside money to pay taxes) the following year. But other costs may remain hidden longer or may never be fully appreciated. For example, vehicle wear and tear may accelerate the need for repairs and replacement, but these costs may not be visible for several years, and drivers might not associate them with their ridesharing activities when they occur.
Finally, we examined broader questions of how sophisticated forum participants are in their understanding of tax law, how they conduct themselves (or, at least, talk about conducting themselves) in terms of tax compliance, and how other cultural characteristics of the discussion boards may shape the transmission of tax knowledge online.
As discussed, we observed a range of sophistication in how well forum participants understood taxes. Discussions on Reddit and Uberpeople seemed more sophisticated than those on TurboTax in several respects. Discussions on Reddit and Uberpeople often used tax terms and concepts more accurately. Participants in the TurboTax forums asked more basic questions about where and how to enter income, whether they had to file taxes, what software to use, and what to do with Form 1099. There were a number of questions that seemed relatively inchoate on TurboTax. Admittedly, however, it was difficult to get a robust sense of poster sophistication on TurboTax because there was much less back-and-forth conversation: participants tended to pose a question, and then a TurboTax employee, professional, or “superuser” would answer that question.
Another interesting feature was the process by which disagreement was expressed and errors corrected on Reddit and Uberpeople. When a poster said something wrong (e.g., took a wrong tax reporting position or gave an inaccurate answer to some else’s question), the error was often, and often quickly, corrected by other posters. For example, in one of the forums, where several posters said they would deduct uniforms and haircuts, others chimed in to say that these items are not generally deductible. When one poster expressed an intent to deduct meals, another immediately countered that meals are not generally deductible. Where one poster opined that taxes would not be owed if grossing less than $20,000, other forum participants were swift to correct the error.
We observed error correction in the other forum as well. When some forum participants suggested that they might not report income because they did not receive a Lyft Form 1099-K, others (including some marketers of tax services) pointed out the audit risk and the obligation to include income. On one discussion thread, the thread originator posted an image of his tax return and the expenses he wrote off, which showed a net loss. The poster reported filing that tax return with the help of a “reputable tax agent.” Other posters were quick to point out problems with his return, making statement such as:
- “$[XXX] for waters and snacks? On just $[XXXX] in gross fares? Gotta say that is some creative accounting.”
- “How did you take into account what were personal expenses vs. business expenses? In other words was all that gas used for Uber only? How could you differentiate it? IRS may want to know in an audit.”
- “I agree with other posters…think you will be in big trouble if audit. at $[X] gal (not sure what you pay) you used [XXX] gallons of gas for only [XXX] trips??? No wonder you drank so much water! I will bet you did not keep track of actual miles, right? Share them if you did. I drove [XXXXX] miles in [X] months!! about [XXXX] was trip miles.”
- “I would be worried about this tax person, and would not recommend him. (I’m a former CPA, by the way.) I would also immediately ask him about a few very obvious issues…”
- “I can tell you that the numbers you have seem way out there to me and I agree completely with the poster who is a CPA that there are some very questionable things going on with your return.”
These comments represent an instance of robust and unequivocal error correction.
Disapproval by Upvoting and Downvoting. A notable feature of Reddit was that even if an error was not explicitly corrected or contradicted, disapproval or disagreement sometimes occurred by upvoting and downvoting. For example, where one poster stated, with respect to whether “dead miles” count as business mileage, “If you drive around looking for fares that’s a valid business purpose. Show me where it’s not,” that statement received 8 points (upvotes minus downvotes). When someone else responded “Show me where it is. You are not actively engaging in commerce unloaded,” that comment received –6 points (i.e., six more downvotes than upvotes). When another poster wrote “I think we should simply not pay the IRS anything. Fuck them. Fuck the United States. This country can seriously choke on a gay cock,” that statement received –4 points. (In the latter instance, it is unclear whether the downvotes were triggered by the tax non-compliance or the phrasing.)
Thus, on Reddit, there was arguably more than one mode of error correction, and we speculate that, among some Reddit users, upvoting and downvoting might be another powerful signal of community sentiment, in addition to outright disagreement. But upvoting and downvoting is not a perfect signal of accuracy. We observed instances where Reddit users sometimes upvoted popular but not totally accurate advice. For example, where someone wrote, “They only track your miles while on a fare. It’s extremely important to track ALL of your mileage, as it is all deductible,” that comment, while not necessarily completely accurate, received 5 points.
In sum, we observed that error correction was surprisingly robust and frequent. This created an overall impression of a compliant culture on the threads, at least in the case of correcting egregious errors. Error correction, however, might be less a matter of compliance and more a matter of asserting and maintaining status and respect in an online community. Error correction was, of course, an imperfect process. It did not always happen, and even when it did happen was not always effective. We observed instances in which errors, misstatements or misunderstandings of the law, or expressed intent not to comply went uncorrected. We also observed instances where, even if corrected, corrections were subsequently undermined by recalcitrant commenters who perpetuated the error. For example, on one thread, accurate commenter statements of which miles are deductible were drowned out by repeated restatements of more aggressive/wrong tax positions by other commenters, such that by the end of the thread, it may have been impossible for the lay reader to tell which was the correct rule. Finally, in all of this, it is unclear whether posters were compliant or accurate in their actual tax behaviors, notwithstanding the existence of error correction on the boards. However, to the extent forum readers may be influenced by their perceptions of the conduct and norms of their peers, the tenor of the forums would likely be influential.
Related to the question of how errors get corrected in the forums is the question of how tax advice is given and received. As our discussion above shows, there was a notable amount of accurate tax advice and commentary given on the discussion boards. In many instances, forum participants on Reddit and Uberpeople stated the tax law accurately, correctly characterizing, for example, the rules with respect to which miles count, whether certain items can be deducted, whether income had to be reported, and where on the tax return to report it. Forum participants also often referred directly to IRS websites and publications in formulating advice. Several forum participants also shared spreadsheet and third party applications that they used in filing taxes, documenting mileage, etc.
However, instances of accurate tax advising were interspersed with inaccurate, vague, or misleading statements about the tax law. As a consequence, while we were able to identify good advice when it occurred, we also took note of many instances of oversimplified, confusing, or just wrong tax advice. To take just a few examples:
- “Be careful, if you lose money 3 out of 5 years, you become not-for-profit and cannot count those losses against other income.”
- “If you don’t make a[ny] taxable income, then your business is considered a HOBBY by the IRS and every penny is taxable.”
- “If the time comes when you get audited, you can create your own documentation at any time.”
- “You cannot deduct .56 cents per mile if your vehicle generates income from even an ‘informal’ taxi service. You must deduct actual expenses and depreciaton [sic] incurred, unless you are leasing your vehicle.”
- “I deduct one meal a day, it’s been awhile since I looked it up. Just don’t eat at a fancy restaurant.”
- “This is what I do and I don’t know if it’s going to work. I know my average MPG is almost exactly 30, so I just take that and look at my receipts and find the total gallons I bought. 30×11 gallons for example = 330 miles. Boom done.”
- “Once a month my bride and I go out for a decent meal out to discuss such topics as Uber, our rental properties, and how I’m running my consulting business. We also talk about our kids, what’s happening on Veep, and where we want to vacation next summer. I’ve been expensing these outings for better than 10 years. Never been a problem. Just make sure you hold onto receipts and don’t abuse the privilege.”
Given that good advice was intermingled with vague, misleading, or inaccurate statements, it is difficult to know whether readers of the discussion boards were able to appreciate and understand the good advice, or the extent to which they were misled by bad advice drowning out the good.
We also observed a fair amount of cautioning or disclaiming that accompanied the tax advice dispensed on the discussion boards. For example, forum participants urged each other to consult with professional tax advisors, and told each other not to rely on the tax advice they themselves gave on the boards. Posters frequently referred each other to IRS websites and publications (either by linking to them or by excerpting and quoting text) in supporting statements they were making. We wondered whether this might either reflect some sense of responsibility for the dispensing of tax advice, a perception that there was some risk or consequence to themselves (either legal or social) from dispensing inaccurate advice to other forum readers, or some other community norm in the forums. We have no idea whether such disclaimers were effective in terms of encouraging readers to seek professional tax help.
In addition to ad hoc comments and advice by various posters, we also observed instances of CPAs, former CPAs, and other self-identified experts answering or volunteering to answer questions. Some of these posters used the forums as a platform to market their businesses or web and mobile applications they had developed. Others just seemed to be offering free advice and sharing their experience, expertise, and spreadsheets, though perhaps with the anticipation of future profit.
The most prominent example of “marketed” tax advice in both the Reddit and Uberpeople forums was the poster USER176/ USER153. This poster was the owner of the now-defunct tryzen99.com website. He originated 7 separate threads on Reddit and commented 63 times, making him the sixth most frequent commenter. On Uberpeople, he originated three threads and commented 67 times. The threads he originated pertained to receipt of Forms 1099 from Uber and Lyft, how to read the forms, other issues confronting ridesharing drivers, and general issues relevant to independent contractors (such as best cities for independent contractors, and health insurance). This poster’s comments spanned several issues, including which miles “count” as deductible business miles, how to understand the Forms 1099-K received from Uber, deductibility of items such as meals, choice of method (standard mileage vs. actual costs), and other issues facing independent contractors. He also posted a number of links to his website, blog, and ridesharing tax guides he had published. He would sometimes interject in discussion threads posted by other marketers to steer the conversation in different directions.
Overall, we found that USER176/USER153’s advice was in the ballpark of accurate, but there were some problem areas. For example, at one point he claimed that drivers could only deduct 50% of water or snacks provided to passengers. But that advice was subsequently recanted in the ridesharing guide published on tryzen99’s website. In addition, prior to January 2015, he advised readers that they would not receive Forms 1099-K unless they met the 200 rides/$20,000 threshold. This turned out not to be the case, as Uber decided to issue Forms 1099-K to all drivers.
Perhaps most importantly, USER176/USER153 consistently took the position in one of the forums that all miles driven on Uber business, including miles driven from home to one’s area of work, counted as deductible business miles. As we have noted, this is a fairly aggressive position and it is possible that the commute miles (i.e., miles driven from home to area of work) may not be written off, unless driving to a different metropolitan area. Other forum posters challenged him on this point but he persisted in taking this position. Interestingly, USER176/USER153 is far less specific in his description of what miles count as business miles in his guide published on the tryzen99 website. In addition, his posts in the other forum were much more equivocal and less aggressive about whether commute miles can be deducted; here, he cautioned against deducting miles from home to area of work, and acknowledged that this is an open issue.
Finally, we noted that even where the advice given by USER176/USER153 was not outright wrong, the nature of posting in an internet forum meant that his advice was sometimes given in overly general terms and could appear ambiguous or confusing. For example, in one case, he gave advice about deducting mileage but did not adequately distinguish between the standard mileage and actual costs methods. We know that he was correctly advising forum readers of the difference between the two methods in other posts, but taken out of context, his statements in this instance could be interpreted as inaccurate. Ultimately, we have no way of knowing how the incidence of error and the quality of tax advice on the boards compares to in-person peer advising or tax preparer advising.
Apart from USER176/USER153, we also counted numerous other instances on both Reddit and Uberpeople of marketed tax advice and promotion of applications and websites designed to help with tax compliance. Forum participants posted tax and ridesharing guides, links to advising websites, and mobile applications that they were building to help with tasks like expense or mileage tracking. These posters sometimes requested that other participants test or fund such applications. On one thread, a poster professing to be a tax attorney offered to answer general questions, and even provided a Circular 230 disclosure.
The internet forums that we studied contained a treasure trove of information about how ridesharing drivers encountered the tax system and interacted with each other online, yielding a rich set of insights. Three notable motifs were apparent from the data: (1) forum posters’ struggles with tax compliance issues; (2) ongoing questions about whether driving was profitable or made economic sense (and how this analysis changed over time); and (3) the unique nature of communication, culture, and community in the internet forums. As this Part discusses, each of these motifs suggests a particular set of normative insights, which may be in tension with each other.
One of the most dominant themes we observed running through the forums was that of expense taking and its attendant difficulties. From calculating mileage to determining which miles are deductible business miles to deciding which expenses are deductible, the question of how and whether to properly recover expenses was a recurring refrain. Relatedly, the question of how to properly document and keep track of expenses was a key challenge for forum participants, as was the question of how to understand and interpret numbers on Form 1099-K when filling out the tax return. In these questions, we saw a good deal of confusion and disagreement regarding the proper application of tax law to ridesharing. For example, forum participants disagreed about whether “dead miles” could be deducted, which apps suffice for documentation purposes, and which expenses could be deducted.
The debates we observed suggest that inaccurate expense taking may be occurring in the ridesharing sector, a result roughly consistent with findings in existing tax compliance literature. Thus, IRS clarification in this area could be a helpful and realistic option. As a first step, it would be relatively simple for the IRS to clarify how the existing rules for deductibility of commuting and other miles apply by extension to ridesharing and to clarify which internet and mobile applications are acceptable for documentation purposes. It would also be easy to issue guidance on how the Form 1099-K rules apply to the ridesharing (and other sharing economy) platforms by clarifying, for example, whether the Form 1099-K reporting positions taken by Uber and Lyft are correct, both with respect to choice of form and also choice of reporting threshold. Guidance targeted at sharing economy workers has already been issued or is being considered by other jurisdictions. As this article was going to press, the IRS began to take initial steps in this direction by creating a webpage for workers in the sharing economy with links to generalized guidance on topics such as business expenses and employments taxes. This is a useful first step. However, the more targeted efforts detailed above would be invaluable in achieving clarity in the tax law and accuracy in ridesharing drivers’ compliance.
Assuming that Uber’s Form 1099-K reporting position is correct, the IRS might also perform driver education—or require the platforms to educate drivers—regarding the need to deduct on their tax return Uber’s commission and other expenses from the gross income number reported on Form 1099-K. As discussed, conversations on the discussion boards raise the possibility that some first-time Form 1099-K recipients may actually have over-reported earnings on their tax returns because Uber reported gross earnings on that form, including tolls, fees, and the Uber commission. The correct approach, as noted by various posters, is for drivers to report the gross income amounts from Form 1099-K on Schedule C of their tax return and then total up tolls and fees (including the split rides fee, the safe rides fee, the Uber commission, and any device subscription fee) and deduct them on Schedule C Line 10 (Commissions and Fees). But this exercise is complicated. The tolls and fees that must be subtracted are not listed on Form 1099-K itself but must be looked up separately on the year-end Tax Summary sent by Uber to each driver. To even discover the gross nature of the Form 1099-K numbers, drivers would have to compare two documents—Form 1099-K and the Uber Tax Summary. While savvier posters seemed to eventually get it right, we do not know whether less savvy ones did. Comparable gross-basis (i.e., fee and commission inclusive) reporting positions may have been or may be taken in the future by other sharing economy platforms, with similar consequences.
Part of the unstated problem is that rideshare drivers may conceptualize themselves as receiving “$100 dollars less 20%” at the outset from Uber, rather than perceiving themselves as independent business owners receiving a gross payment ($100) and then having to pay an expense (the Uber commission and fee) to earn that amount. Thus, in part because ridesharing drivers may not see themselves as operating a small business, gross-basis information reporting may be particularly dissonant in this sector and may give rise to tax reporting mistakes. Clarification or driver education—whether done by the IRS or by the platforms—may help alleviate confusion.
Content aside, there are also different methods by which the IRS might intervene, advise, or clarify. A less invasive approach would be to issue regular publications or statements on the IRS website. A more invasive approach might be to directly intervene on the discussion boards by having a representative post content in the forums. A possible middle ground would be to use social media channels, such as the IRS’s existing Twitter, Tumblr, and Facebook pages, to do driver outreach. As further discussed in Part VI.C, each of these methods may reverberate through internet communities in different ways. Moreover, it is worth noting that the success of some of these methods will depend in part on the sophistication of tax conversations in the forums. For example, the efficacy of IRS social media postings will depend in part on whether forum posters raise relevant questions and share such postings with other forum participants.
Another important theme was the question of whether driving was profitable and how to evaluate profitability. Even though we had initially focused on tax-specific themes, we saw many discussions of how various non-tax expenses—including gas, water, car washes, costs of cleaning up vomit, and vehicle wear and tear—affected profitability. Forum posters also discussed how trends in profitability (as affected by surges, guarantees, and changes in base rates) affected their driving decisions and expressed significant anger and frustration at declines in profitability. They often compared the money they would make driving with the likely hourly rate they would earn at low-wage jobs such as Walmart.
Taxes entered these discussions as a component of driving expenses, which would reduce gross fares earned. The IRS standard mileage rate also served as a proxy for what posters thought the expenses of driving might be, though there was disagreement about the accuracy of this proxy. In these discussions, we saw confusion, guesstimation, and disagreement over how to evaluate profitability and how to estimate the ultimate tax burden. There were significant differences in how various posters estimated driving expenses, anticipated federal and state tax burdens, and made provisions for payment of taxes (e.g., by setting aside money to pay taxes or increasing withholding on W-2 income). As noted, some of these behaviors could be consistent with the spotlighting and ironing hypotheses of the salience literature.
Thus, although we had not explicitly sought to investigate the topic of money and profit, this was a resoundingly pervasive theme in the forums that suggests several normative insights. First, there is the question of whether there should be outreach or driver education with respect to questions of profitability. The rideshare industry is one in which many of the operating costs, and the burdens of estimating such costs, are borne by drivers. The online conversation suggests that some of these costs (including tax costs) may not be particularly salient to drivers at the time of their labor supply decision. If this is so, then driver outreach might be advisable in the face of a new business model that shifts hidden costs to drivers. Any such driver awareness or education initiatives could take different forms. For example, the platforms might be required to disclose estimated costs of driving to drivers. Alternatively, outreach might be conducted by government agencies to help drivers understand the true costs of driving. Incidence and magnitude of driver costs should also be carefully considered in the current debates over worker classification: Legislators, courts, and regulators might consider how employee classification—which would mean wage withholding and other non-tax benefits but also more limited ability to deduct certain business expenses for tax purposes—might play out in an industry where the bulk of costs would likely remain on drivers.
More broadly, if driver sophistication in assessing profitability were to increase over time, this could affect Uber’s business model going forward. Uber’s model is predicated on attracting a large base of drivers and passengers to the platform. If drivers change their driving behaviors as they become more aware of expenses and tax burdens, this realization may affect the viability of the core business model. If, however, drivers do not change their labor supply choices over time, their continued willingness to drive may suggest that the business model is actually profitable for them. But it may also suggest that drivers’ ability to accurately assess profitability remains limited and that more robust driver protections are necessary.
A third theme that ran through the data was the nature of communication and culture in the online communities we studied. The tax talk contained on the discussion boards was a messy conversation that went back and forth and was sometimes rude, pointed, and angry but was at other times supportive and communal. Forum users corrected each other’s errors (sometimes imperfectly); gave each other advice (much accurate, some inaccurate); expressed approval or disapproval in various ways (explicitly, as well as by upvoting and downvoting on Reddit); argued with each other; and liberally shared information about their tax return preparation, documentation, fudging, and other practices. Forum posters marketing commercial interests, suspected plants, members of the press, researchers, taxi drivers, and other interlopers interjected in these conversations, and a few (like USER176/USER153) became recognized and familiar contributors on the boards. At the same time, Reddit and Uberpeople are communities with porous boundaries. While there were some repeat posters, there were many infrequent users and likely an even larger number of lurkers. Forum posters seemed aware that there was a possibility that others (including Uber employees, non-drivers, and the IRS) were reading the forums and, in fact, constantly discussed and referred to the presence of these outsiders, but worries about information being read by outsiders did not seem immensely salient. While we saw a few subsequently deleted posts, they were a minority.
The existence of active online communities of Uber drivers who discuss and debate tax and other issues surrounding rideshare driving is unsurprising in a world with widely available technology, a free speech ethos, and challenging tax obligations. Thus, the issue is not whether such communities and conversations should exist, but rather whether and in what ways we might imagine influencing or harnessing them. In considering such interventions, the distinctive characteristics of the forums (including the variety of participants, accuracy of advice, and methods of error correction) raise questions and suggest that there are tensions that need to be managed: First, there clearly is a process of learning about tax law that is taking place in the forums through conversations and advising among peers and through correction of each other’s errors. This peer-to-peer learning and advising process may fill a need that is currently unmet by the tax authority or by professional tax advisors. Second, however, this learning process occurs in fits and starts and may be muddied and interrupted by error propagation due to advice that is confusing, poorly stated, or just wrong. Third, self-proclaimed professionals or experts, who may be trying to grow their businesses or may be seeking status due to claimed expertise, also give advice in these forums. Fourth, this whole messy conversation can be read by anyone with internet access. And finally, with the exception of a few marketers operating under their real identities, there is little likelihood of imposing ex post consequences on posters through lawsuits or other non-legal means because of poster anonymity. While we did observe “soft” peer sanctions in the form of error correction or telling each other off, it is unclear whether these were a sufficient consequence to effectively discipline or regulate the conversation in terms of accuracy.
Any attempts to intervene in or regulate peer legal advising on the internet boards would need to consider and manage these tensions. With respect to regulation of peer advising, for example, it is possible that such regulation is not advisable because the gains from peer learning on the forums outweigh the downside of confusion or error propagation. Of course, it is difficult to know if this is actually the case, especially once positive and negative externalities on lurkers are considered. Another possibility is that marketers and other professionals ought to be subject to regulation while peer advisors should not. However, such an approach would create obvious line-drawing issues.
Outright regulation aside, other interventions or monitoring of the discussion forums by a taxing authority would also likely have impacts that should be carefully considered. As we suggested in Part VI.A, for example, the IRS might issue guidance with respect to tax issues concerning rideshare drivers and might disseminate such information in online forums where drivers assemble. The IRS could also monitor the forums to obtain information about emerging tax compliance issues, or it could intervene directly in internet forums to correct error propagation and stake out agency positions. Each of these approaches raises questions about what the impacts of such interventions might be on internet communities. For example, if it became apparent that the IRS was monitoring discussion boards to keep apprised of likely major tax compliance issues, this knowledge might have a chilling effect on discussions and might compromise the openness of information sharing on the boards, disrupting the peer learning process. If, instead, the IRS was actively engaging and answering questions on internet forums (e.g., “Q&A Hour with the IRS!”), this may encourage more open discussion, but it could also expose the IRS to risk if advice is misinterpreted or misperceived.
Development of a full theory of regulatory intervention on internet discussion boards is a project for another day. We note, for now, that there are promising ways for taxing authorities to use information discussed online, and to intervene and do taxpayer outreach on internet forums and via social media. The use of social media for tax administration in the sharing economy is also being considered in other countries as well. In designing such outreach, a variety of formats might be explored, with some approaches more interactive than others. For example, running a website that allows users to pose questions and then selecting which questions to answer is an example of a less interactive approach that might nonetheless be effective. Ultimately, the issue is one of design choice, in which potential costs, benefits, and tradeoffs would need to be balanced.
In this Article, we investigated the tax issues and challenges faced by ridesharing drivers by analyzing their interactions in three internet discussion forums. Our detailed study of user interactions in these internet forums is the first of its kind in the legal literature. It offers a timely and nuanced picture of how the nascent encounter between ridesharing workers and the tax regime has played out “in action” on internet discussion boards.
We presented a series of findings about the tax compliance challenges encountered by ridesharing workers, examined the ways taxes factor into their profitability determinations, and discussed cultural aspects of how tax advising and information exchange operate in these internet discussion forums. Based on these findings, we made normative recommendations for more effective tax administration, spelled out the implications of our study for Uber’s business model and for potential regulation of ridesharing, and discussed the potential impacts of various tax compliance initiatives on online communities of drivers and other 1099 workers. These recommendations may be in tension. For example, the need for taxpayer education, which may be accomplished through online forums, raises the risk that IRS intervention in the forums may change the nature of community and communication currently present in such forums. Therefore, any such interventions would need to be carefully considered in order to arrive at the optimal policy choice.
In addition to yielding specific insights concerning ridesharing drivers, our study also has potentially broader application to other gig economy workers. Careful analysis of internet forum discussions can provide much needed texture to our understanding of how 1099 workers deal with taxes and interact with the tax system and how taxes factor into their decision making. Thus, our analysis provides a roadmap for identifying issues that may be important in the broader 1099 economy and may merit more in-depth research.
Acknowledgements: We are grateful to the participants of the University of Virginia Invitational Tax Conference, the Duke Law School Tax Policy Workshop, the Indiana University Maurer School of Law Tax Policy Colloquium, the Georgetown University Law Center Tax Law and Public Finance Workshop, the Brooklyn Law School Faculty Workshop, the Pepperdine School of Law Tax Policy Workshop, the UC Irvine School of Law Tax Policy Colloquium, the Southwestern Law School Faculty Colloquium, the Tulane Tax Roundtable, and the 2015 National Tax Association Annual Conference for helpful comments on this article. Our particular thanks to James Alm, Jake Brooks, Paul Caron, Allison Christians, Steven Dean, Michael Doran, Lilian Faulhaber, Adam Feibelman, Pamela Foohey, Brian Galle, David Gamage, Sara Greene, Itai Grinberg, Andrew Hayashi, Kimberly Krawiec, Rebecca Kysar, Sarah Lawsky, Leandra Lederman, Ann Lipton, Omri Marian, Ruth Mason, Jacob Nussim, Leigh Osofsky, Mildred Robinson, Adam Rosenzweig, Emily Satterthwaite, Darien Shanske, Stephen Shay, David Walker, Ethan Yale, George Yin, and Lawrence Zelenak for their feedback on drafts. The authors would like to thank Alice Huang and Kyle Liftin for excellent research assistance, and would also like to thank the Dr. Thomas F. Carney ’47 Gift Fund at Boston College Law School for its generous support.
Appendix: List of Codes
ATTITUDES TOWARDS TAX COMPLIANCE
SUBSTANTIVE TAX ISSUES
Employee or Independent Contractor
Entities and Structuring
Expenses and Deductibility
Form 1099: Content and Accuracy
Form 1099: Different Reporting Positions
Form 1099: Receipt
Form 1099: Residual
Interest and Penalties
IRS Website and Publications
Method: Standard Mileage or Other
Tax Prep Software
Tax Preparer or Advisor
Third Party Apps
Water and Snacks
STRATEGIES AND RATIONALES
“Beer money” vs livelihood
New Year’s Eve
Profitability: Before and After Tax
Profitability: Decision to Drive
Profitability: Per Mile vs. Per Hour
TAX POLICY ISSUES
Fairness and Equity
Sympathy or Commiseration
Link to External Website
Link to External Website
Link to External Website
Unanswered Detail Request
 Uber’s Founding, Uber, http://newsroom.uber.com/ubers-founding/ [http://perma.cc/L38C-V3EN].
 Jonathan V. Hall & Alan B. Krueger, An Analysis of the Labor Market for Uber’s Driver-Partners in the United States (NBER Working Paper No. 22843), http://www.nber.org/papers/w22843 [http://perma.cc/3PU6-JC62]. “Actively partnering” is defined as the provision of at least four trips via UberX or UberBLACK to passengers in a given month. Id. at 2, 6 n.7. According to the study, the rate of growth in drivers signing up to drive is rising exponentially. Id. at 2.
 Heather Somerville, Uber reaches 2 billion rides six months after hitting its first billion, Reuters (July 18, 2016, 11:06 AM), http://www.reuters.com/article/us-uber-rides-idUSKCN0ZY1T8 [http://perma.cc/MVE3-6PEB].
 See, e.g., Airpnp, Facebook, http://www.facebook.com/weallpee/ [http://perma.cc/MF3F-QXQ7] (last visited Sept. 24, 2016) (toilet sharing); AirBnb, http://www.airbnb.com/ [http://perma.cc/33DP-YHF7] (last visited Sept. 24, 2016) (home sharing); Fon, http://corp.fon.com/en [http://perma.cc/BS63-MJDE] (last visited Sept. 24, 2016) (Wi-Fi sharing); Sarah Haydu, Clear Your Calendars—#UberKITTENS Are Back, Uber, http://newsroom.uber.com/uberkittens-are-back/ [http://perma.cc/4UKY-ZKS3] (last visited Sept. 24, 2016) (kittens).
 See, e.g., Jordan M. Barry & Paul L. Caron, Tax Regulation, Transportation Innovation, and the Sharing Economy, 81 Chi. L. Rev. Dialogue 69 (2015); Christopher Koopman et al, The Sharing Economy and Consumer Protection Regulation: The Case for Policy Change, 8 J. Bus., Entrepreneurship & L. 529 (2015); Stephen Miller, First Principles for Regulating the Sharing Economy (Feb. 20, 2015), 53 Harv. J. of Legis. (forthcoming 2016); Stephen Miller, Transferable Sharing Rights: A Theoretical Model for Regulating Airbnb and the Short-Term Rental Market (Oct. 24, 2014) (unpublished manuscript), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2514178 [http://perma.cc/C9H4-X6N2]; Shu-Yi Oei & Diane M. Ring, Can Sharing Be Taxed?, 93 Wash. U. L. Rev. 989 (2016); Sofia Ranchordas, Does Sharing Mean Caring? Regulating Innovation in the Sharing Economy, 16 Minn. J.L. Sci. & Tech. 413 (2015); Sofia Ranchordas, Innovation-Friendly Regulation: The Sunset of Regulation, the Sunrise of Innovation, 55 Jurimetrics 201 (2015); Daniel E. Rauch & David Schleicher, Like Uber, But for Local Government Policy: The Future of Local Regulation of the ‘Sharing Economy,’ 76 Ohio St. L.J. 901 (2015); Kellen Zale, Sharing Property, 87 U. Colo. L. Rev. 501 (2016); Brishen Rogers, The Social Costs of Uber, 82 U. Chi. L. Rev. Dialogue 85 (2015); Abbey Steimer, Betwix and Between: Regulating the Sharing Economy (Ind. Univ. Kelley Sch. of Bus. Research Paper No. 15-6, 2014), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2535656 [http://perma.cc/2UYH-NHMH].
 See, e.g., Eli Dourado & Christopher Koopman, Evaluating the Growth of the 1099 Workforce Mercantus ctr., George Mason U. (2015), http://www.mercatus.org/system/files/Evaluating-Growth-1099_Dourado_MOP_v2.pdf [http://perma.cc/5V3Z-W35R]; Ian Hathaway, The Gig Economy is Real If You Know Where to Look, Harv. Bus. Rev. (2015), http://hbr.org/2015/08/the-gig-economy-is-real-if-you-know-where-to-look [http://perma.cc/GNF4-WYZP]; Justin Fox, The Mystery of the Multiple Jobholder, Bloomberg View (June 26, 2015), http://www.bloombergview.com/articles/2015-06-26/the-mystery-of-the-multiple-jobholder [http://perma.cc/QQ6D-DR7S].
 “Sharing economy” is the term most commonly used, but does not capture the commercial nature of the exchange. The same applies to the term “collaborative consumption.” Alternative terms such as “1099 economy” or “gig economy” have been urged as replacements, but no single usage has prevailed. This Article generally uses the terms “1099 economy” and “gig economy” to describe the sector.
 See, e.g., Brishen Rogers, Employment as a Legal Concept (Temple Univ. Leg. Stud. Research Paper No. 2015-33), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2641305 [http://perma.cc/NBU3-9FFR]; see also The 2015 1099 Economy Workforce Report (May 20, 2015).
 Cotter v. Lyft, Inc., 176 F. Supp. 3d 930 (N.D. Cal. 2016) (for full docket see Docket No. 3:13-cv-04065); O’Connor v. Uber Technologies, Inc., 58 F. Supp. 3d 989 (N.D. Cal. 2014) (for full docket see Docket No. 3:13-cv-03826); Berwick v. Uber Technologies, Inc., Cal. Superior Ct. No. 15-546378 (Nov. 22, 2016) (order of dismissal with prejudice following parties’ settlement) available at http://query.sftc.org/Minds_asp_pdf/Viewer/DownLoadDocument.asp?PGCNT=0; Alatraqchi v. Uber Technologies, Inc., Case No. A142059 (Cal. Ct. App. 2014) (dismissed plaintiff appeal on June 27, 2014; for full docket see http://appellatecases.courtinfo.ca.gov/search/case/dockets.cfm?dist=1&doc_id=2079405&doc_no=A142059 [http://perma.cc/J794-RTFV]). For Fedex and taxicab litigations, see Slayman v. Fedex, Inc., 765 F.3d 1033, 1042-1046 (9th Cir. 2014); Sebago v. Boston Cab Dispatch, Inc., 471 Mass. 321, 327-328 (Mass. 2015).
 For an interview and survey-based study, see Caroline Bruckner, Shortchanged: The Tax Compliance Challenges of Small Business Operators Driving the On-Demand Platform Economy (Kogod Tax Pol’y Ctr., American Univ. 2016) available at http://www.american.edu/kogod/news/upload/shortchanged-caroline-bruckner-kogod-au.pdf [http://perma.cc/Q6M3-ES8B].
 See generally Tom Perez, Innovation and the Contingent Workforce, U.S. Dep’t of Labor Blog (Jan. 25, 2016), http://blog.dol.gov/2016/01/25/innovation-and-the-contingent-workforce/ [http://perma.cc/PR6W-DFQC] (Bureau of Labor Statistics and Census Bureau plans to study contingent workforce as part of May 2017 Current Population Survey).
 See sources cited infra note 53.
 See discussion infra Part VI.A.
Sharing Economy Tax Center, Internal Revenue Service, http://www.irs.gov/businesses/small-businesses-self-employed/sharing-economy-tax-center [http://perma.cc/NGU7-DBSF].
 See discussion infra Part VI.B.
 See discussion infra Part VI.C.
 For a non-legal, sociological study of sharing economy labor that investigates online forums, see Alex Rosenblat & Luke Stark, Uber’s Drivers: Information Asymmetries and Control in Dynamic Work (Oct. 15, 2015) (unpublished manuscript), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2686227 [http://perma.cc/Q6D6-GRVK]. For an example of tax scholarship citing internet discussions for anecdotal evidence of tax compliance behaviors, see Bridget Crawford, Taxing Surrogacy, in Challenging Gender Inequality in Fiscal Policy Making: Comparative Research on Taxation, at 95-108 (Asa Gunnarson et al., eds. 2011); see also Jeffrey Hoopes et al., Taxpayer Search for Information: Implications for Rational Attention, 7 Am. Econ. J. 177 (2015). For examples of scholarship in other fields, see sources cited infra note 53.
 See generally Roscoe Pound, Law in Books and Law in Action, 44 Am. L. Rev. 12 (1910).
 Uber, http://www.uber.com/ [http://perma.cc/9W8Z-593G] (last visited Nov. 20, 2016); Erin Griffith, Uber v. Lyft: The Credit Cards Don’t Lie, Fortune (Sept. 11, 2014, 9:00 AM), http://fortune.com/2014/09/11/uber-vs-lyft-the-credit-cards-dont-lie/ [http://perma.cc/RUZ7-WY93].
 See Uberestimator, Cities, http://uberestimator.com/cities [http://perma.cc/QJ5F-F5UK] (last visited Nov. 20, 2016); Uber, supra note 19.
 How does Uber work?, Uber, http://help.uber.com/h/738d1ff7-5fe0-4383-b34c-4a2480efd71e [http://perma.cc/U7DL-D46L] (last visited Sept. 24, 2016).
 See Alex Wilhem, In Another Strike Against The Competition, Uber Lowers UberX Prices in San Diego, LA, and DC, Tech Crunch (Oct. 3, 2013), http://techcrunch.com/2013/10/03/in-another-strike-against-the-competition-uber-lowers-uberx-prices-in-san-diego-la-and-dc/ [http://perma.cc/C4KB-T7DC].
 Uber, http://newsroom.uber.com/drive-with-uber-earn-cash-with-your-car-6/ [http://perma.cc/K35W-MLWS] (last visited Nov. 20, 2016).
 See generally Uber, supra note 19.
 What is Surge?, Uber, http://help.uber.com/h/e9375d5e-917b-4bc5-8142-23b89a440eec [http://perma.cc/6WFQ-UZEY] (last visited Dec. 22, 2016).
 See Nicholas Diakopoulos, How Uber surge pricing really works, Wash. Post (Apr. 17, 2015), http://www.washingtonpost.com/news/wonkblog/wp/2015/04/17/how-uber-surge-pricing-really-works/ [http://perma.cc/F7DF-W554].
 Uber, http://help.uber.com/h/f7385bf5-1748-4fd0-a57f-3d9b62facc45 [http://perma.cc/5EF2-M6G8]; Jay Barmann, Now Uber Drivers Want You To Tip, Sfist (Feb. 2. 2015), http://sfist.com/2015/02/17/now_uber_drivers_want_you_to_tip.php [http://perma.cc/W8R6-7RFL]. This tip policy has created some dissatisfaction among drivers, leading to litigation. See Cotter v. Lyft, Inc., 176 F. Supp. 3d 930 (N.D. Cal. 2013) (for full docket, see Docket No. 3:13-cv-04065); O’Connor v. Uber Technologies, Inc., 58 F. Supp. 3d 989 (N.D. Cal. 2014) (for full docket, see Docket No. 3:13-cv-03826); Maya Kosoff, Uber’s Drivers Say They Don’t Get Any Tip Money From All-Inclusive Fares — And They’re Furious, Bus. Insider (Sept. 7, 2014 at 3:09pm), http://www.businessinsider.com/ubers-drivers-say-they-dont-get-any-tip-money-from-all-inclusive-fares-2014-9 [http://perma.cc/LD33-UPW9].
 Douglas MacMillan, Uber Tests 30% Fee, Its Highest Yet, Wall St. J. (May 15, 2015), http://www.wsj.com/articles/uber-tests-30-fee-its-highest-yet-1431989126 [http://perma.cc/DG3Y-ZT82]. See also Uber, http://uberestimator.com/booking-fee [http://perma.cc/4C44-TX97] (explaining safe rides fee). The safe rides fee has been the subject of litigation. Philliben et al v. Uber Technologies, Inc., Case No. 3:14-cv-05615, 2016 WL 6840229 (N.D. Cal. 2016). See also Ellen Huet, Uber Faces Class-Action Lawsuit Over $1 ‘Safe Rides Fee’, Forbes (Dec. 27, 2014), http://www.forbes.com/sites/ellenhuet/2014/12/27/uber-class-action-lawsuit-safe-rides-fee/ [http://perma.cc/9Z7G-LXR3].
 Caren Ehret et al. v. Uber Technologies Inc., 68 F. Supp. 3d 1121(N.D.Cal. 2014) (customer/plaintiff claimed that Uber misled customers by charging 20% tip that does not go exclusively to the driver); see also Joel Rosenblatt, Uber Wins Approval of Settlement with Customers Over Tips, 21 ECLR vol. 37 (Bloomberg BNA, Sept. 21, 2016).
 See Uber, http://help.uber.com/h/ba14dbf4-0e34-452b-ab0c-22f7d65b8d61 [http://perma.cc/VBE5-VFFK] (last visited Dec. 6, 2016); Luz Lazo, Uber gives its drivers a choice to avoid $10 weekly fee for app use, Wash. Post (Sept. 9, 2014), http://www.washingtonpost.com/blogs/dr-gridlock/wp/2014/09/09/uber-gives-its-drivers-choice-to-avoid-10-weekly-fee-for-app-use [http://perma.cc/F7X8-LCWB].
 See Jon Kelly, The Uber vs. Lyft War Has Been Won in Silicon Valley, Vanity Fair (Sept. 3, 2015), http://www.vanityfair.com/news/2015/09/uber-vs-lyft-war-has-been-won [http://perma.cc/CV38-R7GP].
 Farhad Manjoo, Uber and Lyft Have Become Indistinguishable Commodities, N.Y. Times Bits (Aug. 28, 2014 at 2:27 pm), http://bits.blogs.nytimes.com/2014/08/28/uber-and-lyft-have-become-indistinguishable-commodities/?_php=true&_type=blogs&_r=1 [http://perma.cc/XV7Q-A7CK].
 Vivek Saxena, Lyft v. Uber: What’s the Difference Between These Two Dueling Apps, Inquisitr (Aug. 14, 2014), http://www.inquisitr.com/1409677/lyft-vs-uber-whats-the-difference-between-these-dueling-apps/ [http://perma.cc/U4MB-YK4P].
 See, e.g., Lyft, http://help.lyft.com/hc/en-us/articles/214219657-How-to-Give-a-Lyft-Ride [http://perma.cc/5KKV-YDLD] (last visited Dec. 22, 2016).
 Lyft, http://www.lyft.com/ [http://perma.cc/Z6B7-6S49] (last visited Nov. 20, 2016). Lyft has also started offering a “Premier” service, which allows customers to request a high-end vehicle. Lyft, http://help.lyft.com/hc/en-us/articles/222360127-Lyft-Premier-for-Passengers [http://perma.cc/574T-EN75] (last visited Dec. 6, 2016).
 Lyft, http://help.lyft.com/hc/en-us/articles/214586017-Prime-Time-for-Drivers [http://perma.cc/V3BU-5X6U] (last visited Nov. 20, 2016).
 Lyft, http://help.lyft.com/hc/en-us/articles/213584308-Insurance-Policy [http://perma.cc/HZ6F-X7XA] (last visited Dec. 6, 2016).
 Lyft, http://help.lyft.com/hc/en-us/articles/213583978-How-to-Tip-Your-Driver [http://perma.cc/B5WU-UGLF] (last visited Nov. 20, 2016).
 Dana Kerr, To tip or not to tip drivers, that is Uber’s question, CNET Mag (Feb. 16, 2015), http://www.cnet.com/news/to-tip-or-not-to-tip-drivers-that-is-ubers-question/ [http://perma.cc/MMQ8-RGKU].
 Lyft, http://help.lyft.com/hc/en-us/articles/214212387-How-Driver-Pay-Is-Calculated- [http://perma.cc/X3LS-HMJZ ] (last visited Nov. 20, 2016).
 See Douglas MacMillan, Sidecar Technologies Shuts Ride-Sharing and Delivery Service, Wall St. J. (Dec. 29, 2015, 11:39 PM), http://www.wsj.com/articles/sidecar-technologies-shuts-ride-sharing-and-delivery-service-1451450372 [http://perma.cc/8EUW-LV29].
 See, e.g., Carolyn Said, Ride-sharing pioneer Sidecar to shut down ride, delivery service, SFGate (Dec. 29, 2015, 3:02 PM), http://www.sfgate.com/business/article/Ride-sharing-pioneer-Sidecar-to-shut-down-ride-6726144.php [http://perma.cc/7656-4ARS]; Douglas Macmillan, Sidecar Succumbs to Uber and Lyft in Car-Hailing Wars, Wall St. J. (Dec. 29, 2015, 5:35pm), http://blogs.wsj.com/digits/2015/12/29/sidecar-shuts-down-succumbing-to-uber-and-lyft/ [http://perma.cc/W6JY-AEZ4].
 I.R.C. §§ 61, 162.
 Oei & Ring, supra note 5.
 Uber, http://get.uber.com/drive/ [http://perma.cc/6LD5-YWV5]; Uber, http://www.uber.com/legal/usa/terms [http://perma.cc/56K8-PVTC]. The position taken by Uber and Lyft that drivers are independent contractors has not gone uncontested. Developments in class action court cases against these platforms, and rulings by state labor and employment tribunals that drivers are properly classified as employees, have brought a new degree of uncertainty to this determination. See sources cited supra note 9. In 2016, Uber and plaintiffs sought to settle two class actions (one from Massachusetts but transferred to the Northern District of California and one from California). However, on August 18, 2016, the U.S. District Court for the Northern District of California rejected the settlement. Order Denying Plaintiffs’ Motion for Preliminary Approval of Settlement, U.S. District Court for the Northern District of California, August 18, 2016, O’Connor v. Uber Technologies, Inc. et al. (Case No. 13-cv-03826-EMC, Case No. 15-cv-00262-EMC), available at http://www.cand.uscourts.gov/EMC/OConnorvUberTechnologies [http://perma.cc/938E-52B9]. That case has been stayed, pending the outcome of various appeals.
The IRS test for independent contractor status in the tax context is drawn from the general common law test. See, e.g., Independent Contractor (Self-Employed) or Employee?, Internal Revenue Service, http://www.irs.gov/Businesses/Small-Businesses-&-Self-Employed/Independent-Contractor-Self-Employed-or-Employee [http://perma.cc/2D67-RMWC]. See also I.R.C. § 3121(d)(2) (defining employee for employment tax as “any individual who, under the usual common law rules applicable in determining the employer-employee relationship, has the status of an employee”). Although the tax law generally determines worker classification based on the common law analysis, in some circumstances a common law independent contractor will nonetheless be treated as a “statutory” employee for specific provisions of tax law. See, e.g., I.R.C. § 3121(d)(3). Classification of drivers for tax purposes will likely be affected by the outcomes of these court cases and other rulings beyond the tax arena.
 See TurboTax, http://turbotax.intuit.com/tax-tools/tax-tips/Self-Employment-Taxes/Tax-Tips-for-Uber–Lyft–Sidecar-and-other-Car-Sharing-Drivers/INF28820.html [http://perma.cc/8VPJ-8AYW](last visited Dec. 6, 2016).
 See, e.g., Leandra Lederman, Reducing Information Gaps to Reduce the Tax Gap: When is Information Reporting Warranted?, 78 Fordham L. Rev. 1733 (2010) [hereinafter Lederman, Reducing Information Gaps]; Joel Slemrod et al., Does Credit-card Information Reporting Improve Small-business Tax Compliance? (Nat’l Bureau of Econ. Research, Working Paper No. 21412); see generally Leandra Lederman, Statutory Speed Bumps: The Roles Third Parties Play in Tax Compliance, 60 Stan. L. Rev. 695 (2007); James Alm, et al, Do Individuals Comply on Income Not Reported by their Employer? 37(2) Pub. Fin. Rev. 120 (2009).
 Uber and Lyft currently report driver referral bonuses and other special direct payments exceeding $600 on Form 1099-MISC. I.R.C. § 6041.
 I.R.C. § 6050W(e).
 See discussion infra Part V.A.4.
 See, e.g., Aditya Pal et al., Early Detection of Potential Experts in Question Answering Communities, Lecture Notes in Computer Science, 6787, 231-242 (2011); Rosenblat & Stark, supra note 17; Reijo Savolainen, Requesting and Providing Information in Blogs and Internet Discussion Forums, 67 J. Documentation 863 (2011); Winnie Shum & Cynthia Lee, (Im)politeness and Disagreement in Two Hong Kong Internet Discussion Forums, 50 J. Pragmatics, 52-83 (2013); L. Jean Harrison-Walker, E-Complaining: A Content Analysis of an Internet Complaint Forum, 15 J. Marketing Services (5) 397 (2001); Wolfgang Himmel et al., Information Needs and Visitors’ Experience of an Internet Expert Forum on Infertility, 7(2) J. Med. Internet Res. e20 (2005); Jyh-Shen Chiou & Jasi Lee, What Do They Say About “Friends”? A Cross-Cultural Study on Internet Discussion Forum, 24 Computers in Human Behavior 1179 (2008).
 A driver’s financial success may depend, in part, on her ability to manipulate and interpret online data (for example, the ability to identify potential price surges and locations of other drivers).
 Nellie Bowles, The Uber-Loneliness of the Sharing Economy Driver, The Guardian (Feb. 4, 2016), http://www.theguardian.com/technology/2016/feb/04/uber-protest-app-zello-sharing-economy [http://perma.cc/3NWW-FWSM].
 We refer to these discussion forums as Uberpeople, Reddit, and TurboTax.
 See Matt Silverman, Reddit: A Beginner’s Guide, Mashable (June 6, 2012), http://mashable.com/2012/06/06/reddit-for-beginners/ [http://perma.cc/L5BP-6QUL]. See Happy 10th birthday to us! Celebrating the best of 10 years of Reddit, Reddit Blog (June 23, 2015), http://www.redditblog.com/2015/06/happy-10th-birthday-to-us-celebrating.html [http://perma.cc/7R6B-GYNS] (as of June 23, 2015, Reddit boasted 1,715,454,785 comments, 334,626,161 monthly page views, 190,227,552 posts, and 36,136,190 user accounts).
 Silverman, supra note 57.
 Uberdrivers, Reddit, http://www.reddit.com/r/uberdrivers/ [http://perma.cc/8XCF-GP4Q].
 Maeve Duggan & Aaron Smith, 6% of Online Adults are Reddit Users, Pew Res. Ctr. (July 3, 2013), http://www.pewinternet.org/2013/07/03/6-of-online-adults-are-reddit-users/ [http://perma.cc/25ZE-YF8Z].
 Id.; See also Maeve Duggan, It’s a woman’s (social media) world, Pew Res. Ctr. (Sept. 12, 2013), http://www.pewresearch.org/fact-tank/2013/09/12/its-a-womans-social-media-world/ [http://perma.cc/7EEA-U2FW] (“reddit is the only site we’ve measured in which men are significantly more likely than women to be users.”).
 Duggan & Smith, supra note 60.
 Hall & Krueger, supra note 2, at 7. The data in the 2016 study was based on a 2014 web survey completed by 601 drivers (a response rate of about 11%). Id.
 Id. at 7.
 Id. at 8.
 Note that the Reddit statistics are calculated as a percentage of internet users, that is, the survey asked internet users questions about their demographic information and their Reddit use. A conclusion that 8% of male and 4% of female internet users report using Reddit does not mean that two thirds of Reddit users are men. In contrast, the Uber study was a study of Uber drivers themselves and about their demographic information.
 UberPeople.net, http://uberpeople.net/ [http://perma.cc/5E3Y-QUAX]. According to its Facebook page, Uberpeople was founded in April 2014. Uber Driver Forum – UberPeople.NET, Facebook, http://www.facebook.com/Uber-Driver-Forum-UberPeopleNET-1505553926331482/ [http://perma.cc/T7R3-YGV2].
 As of October 2015, Uberpeople had over 26,000 members, over 33,000 discussions and over 486,000 posted messages.
 See Bowles, supra note 55 (calling the site “a sort of public water cooler” and noting that Uberpeople users have made over 750,000 posts to date).
 See UberPeople.Net, Who owns Uberpeople-Net?, http://uberpeople.net/threads/who-owns-uberpeople-net.18569/ [http://perma.cc/A2GV-HK5X].
 See http://uberpeople.net/threads/this-needs-to-be-said-this-forum-is-run-by-and-worked-by-uber-employees.3643/ [http://perma.cc/R8KZ-DVK8] (thread titled “This needs to be said! This forum is run by and worked by Uber employees”).
 TubroTax, http://ttlc.intuit.com/ [http://perma.cc/N895-V8VY] (last visited Dec. 22, 2016).
 Reddit allows wildcard searching using an asterisk after a word. But sometimes wildcard searching was ineffective. For example, a search for tax* yielded threads about “taxis,” and a search for deduct* yielded threads about insurance “deductibles.” In those instances, we did not use wildcards due to over-inclusiveness.
 Other sections included Resources, Members, Blogs, and Cities.
 During this study’s time period, some threads under “Pay” were in neither the “Tips” nor the “Taxes” sub-subheading.
 All threads that contained the word “taxes” somewhere in the discussion were captured. Once the thread was captured, all pages of that thread were saved for study. The word “tax” is not searchable on Uberpeople because it is too short.
 We did not review threads in the “Geographical” category because that category was city specific and did not pertain to taxes at all.
 Atlas.ti, http://atlasti.com/ [http://perma.cc/5JD6-SKEC] (last visited Dec. 22, 2016). In this Article, citations are either to an anonymized list of 177 user handles denoted “USER1, USER2, etc.” or to PDFs of the discussion threads that we analyzed (referred to as “Primary Documents” or “PDOCs”). Some discussion threads spanned more than one PDOC. Thus, there were a total of 487 Uberpeople PDOCs, 291 Reddit PDOCs, 85 PDOCs for TurboTax threads with keyword “Uber,” and 31 PDOCs for TurboTax threads with keyword “Lyft.” For example, “Reddit P1” denotes the first PDOC in Reddit dataset. Usernames and which forum a post came from have been anonymized where appropriate to protect poster identities. The suffix “_TO” denotes a forum user’s thread-originating post.
 Each thread contains the date the comment was posted and the date of the original post.
 Reddit Help, http://reddit.zendesk.com/hc/en-us/articles/204511559-How-is-a-post-s-or-comment-s-score-determined- [http://perma.cc/Y8GK-98GF].
 “Substantive Tax Issues” refers to threads addressing issues of substantive tax law. “Profitability” refers to threads dealing with whether driving for Uber was profitable. “Driving Strategy” refers to threads addressing methods, strategies, and decisions for most effective rideshare driving. “Independent Contractor vs. Employee” refers to threads that deal with issues surrounding worker classification, such as labor law issues or whether drivers should unionize. (Tax issues relating to independent contractor status (such as Form 1099 issues) were grouped in the “Substantive Tax Issues” category.) “Uber’s Business Model” refers to threads pertaining to how Uber conducts its business. “Interactions with Uber” refers to discussions of interactions between drivers and the platform. “Marketing” refers to posts by posters trying to market their software or services. Where thread topics spanned more than one category, we assigned that thread to the category that most accurately reflected its dominant theme.
 We use the terms “handle” and “user” interchangeably.
 This reflects the fact that Uberpeople.net appears to have been started in April 2014.
 This refers to questions about how to prepare the tax return (e.g., where to enter income or deduct fees).
 Our discussion reflects our broad sense of thread impact, based on the metrics (such as number of comments and points scores) discussed in Part IV.
 Additional substantive topics that appeared in Reddit but with less frequency included “Entities and Structuring” (28), “Interest and Penalties” (13), and “State Tax” (39).
 USER6_TO (“so I don’t have to pay taxes if I gross less than $20k?”); USER103_TO (“I only made $[XXXX] for the year. . . . Do I really need to waste time and file taxes over this bullshit? I heard something like if you make less than $10,000 you don’t have to file taxes but I made way more than that at my ‘real’ job.”); USER66 (“If they are not reporting it to the IRS then I’m not going to report it”); USER63 (“do you have to report your earnings from Uber to the IRS, and is there anyway around this?”); USER98 (“Do you think since Lyft didn’t give me form 1099 I should not bother declaring the income from their gig on my 2014 tax returns?”); see also infra note 134.
 USER16; USER84.
 We coded 211 instances of the code “income inclusion” in Uberpeople and 99 instances in Reddit. Interestingly, the code “Income Inclusion” appeared comparatively frequently in TurboTax (102 occurrences, compared with 126 occurrences for “Expenses and Deductibility”), perhaps reflecting the fact that posters were actually filling out tax forms when they posted in this forum.
 See, e.g., USER60 (“Are you implying that Lyft drivers who don’t get a 1099K don’t owe taxes on their earnings? Because that’s not how it works.”).
 See discussion supra Part II.A.2; see also Oei & Ring, supra note 5; Lederman, Reducing Information Gaps, supra note 49, at 1752 (noting that Form 1099-K reporting effectiveness may be harmed by high reporting thresholds, which exclude many taxpayers from reporting); see also I.R.C. § 6050W(e).
 Susan Cleary Morse, Stuart Karlinsky, & Joseph Bankman, Cash Businesses and Tax Evasion, 20 Stan.L.& Pol’y Rev. 37 (2009) (finding, based on field interviews of 275 cash business owners, that most interviewees regarded credit card receipts as taxable and reportable revenue). While it addresses a different group of taxpayers, the study raises the possibility that the electronic nature of amounts earned in ridesharing may incentivize drivers to report such income, regardless of Form 1099-K reporting.
 See, e.g., USER112 (“The water provided for customers is probably OK, but keep the receipts for that, along with any tolls or parking fees”); USER65 (“Car chargers, water and mints for pax…[are deductible]”); but see USER52 (“Gas yes. Water, I don’t think so.”) Other expenses identified by forum participants included business cards, cell phone mounts for the car, and water and mints for passengers. See, e.g., USER135. Drivers discussed the best cars for ridesharing, often with a focus on the vehicle’s expected miles per gallon. Hybrids were frequently cited as a promising choice, at least on cost grounds, if not aesthetic appeal.
 I.R.C. § 274(d); Treas. Reg. § 1.274-5(j)(2); I.R.S. Notice 2014-79 (Sec.3), 2014-2 C.B. 1001; I.R.S. News Release IR-2014-114 (Dec. 10, 2014). Rev. Proc. 2010-51, 2010-2 C.B. 883; Internal Revenue Service Publication 463, “Travel, Entertainment, Gift, and Car Expenses” (2013), 17.
 For example, if the driver uses the actual expenses method and the vehicle were used only 40% of the time (based on mileage) for ridesharing (with the remainder constituting personal use), only 40% of the vehicle expenses would be deductible.
 Rev. Rul. 99-7, 1999-1 C.B. 361. There are three exceptions: (1) If the taxpayer has at least one regular work location away from her home, then commuting expenses between her home and a temporary work location in the same trade or business can be deducted. (2) Transportation expenses for going from home to a temporary work location outside the metropolitan area where the taxpayer lives and works are deductible. (3) If the taxpayer’s residence is her principal place of business, then she may deduct transportation expenses for going between home and another work location in the same trade or business. Id.
 USER98 (“Unfortunately, I didn’t keep log. I didn’t know the rules when I started and Uber doesn’t tell you shit.”); USER102 (“I started fares already without documenting this past week,”). See also the following exchange:
USER10: “Just spoke with a CPA, he said you can count the distance between your drop off point to your next ride. You’ll need to use the uber trip history, trip by trip and find the mileage in google maps.”
USER 23: “Oh god what a pain in the ass that would be.”
USER10: “It’s going to take me weeks”
 USER40 (“We all have proof on the uber dashboard.”); USER43 (“what can i do if i have not been keeping up with miles and gas?”); USER39.
 Uber’s 2014 Tax Summary itself states that “Additional mileage may be deductible.” http://ridesharedashboard.com/2015/01/30/uber-taxes-2015-tax-summary-deducting-fees/ [http://perma.cc/863X-EGF2].
 See, e.g., USER51 (“Uber will only record on your 1099 the number of miles you drove after the pax gets in the car but you can absolutely deduct all business related miles – like driving around to a spot and driving to the pickup point.”); USER132 (“I just thought I could use the miles logged in the waybill, summary, etc statements, but I realize these don’t track your miles spent on picking people up before the ride starts.”).
 USER121_TO (“if you’re not keeping track of every mile you drive, you are screwed because this is where your tax savings will come from”); USER161 (“It’s extremely important to track ALL of your mileage, as it is all deductible.”); USER172 (“Once you get behind the wheel go on line in your driveway you are on the job.”); USER92 (“Turn the app on when you leave your house…, now they are deductible.”); USER61 (“Your place of business is your house, there are no commuting miles.”); USER85 (“I turn the app on before I start the engine. All miles are ‘work’.”); USER147 (“You’re not listening. YES YOU SHOULD DEDUCT ALL THOSE MILES. Your home is your base when you’re an independent contractor. When you leave it, you are travelling for work and should deduct it.”); USER22 (“When you leave your house….. The app is on. Period…. say nothing else to the tax man. When the app is on…. you’re officially working. PERIOD. The only ‘commuting miles’ that you should have as an uber/lyft driver…. are on the soles of your shoes when you walk from your back door to your car.”).
 See supra note 104; see also USER85 (“There is no commute. I am working if the car is moving. There is no set location to commute to. Uber rides begin and end everywhere.”)
 USER32 (This poster notes that they have “made it a point to turn on [the] uber [app] every time [he] get[s] in the car,” which makes all miles deductible. They also note that “[y]ou can always accept rides then cancel if you really have somewhere to be urgently” and that “[t]ime on the clock is time on the clock.”).
 This distinction was most relevant for drivers who, for example, lived outside an active urban area and drove from home to that area to find passengers. For example, consider the following exchange:
USER129: “I’m wondering if you live outside of the Uber coverage area, can you deduct the miles driven from your home to the coverage area? Or, should you consider that ‘commuter’ miles. Any thoughts?”
USER172: “There are people that work in one place regularly and then sometimes have to go to a farther office. They minus out their normal commute from the total trip to the outlying office. With livery you do not have to do that. Once you get behind the wheel go on line in your driveway you are on the job. With the increased insurance coverage you get while you are on duty I would drive to the hot zone being on duty from your home.”
 USER105 (“I’ve heard from my Dad’s tax guy, and also read that: The commute is the first leg, and last leg of your work period. When you first leave to get to where you hang out, and when you are on your way back home are concidered [sic] your commute. Between fares is still working.”); USER98 (“H&R Block tax pros specifically answered this question contrary to what you are saying. NO, you cannot deduct commute miles. This is verbatim what they said: ‘You, an IC, or selfemployed driver are no different from a W2 employee commutting [sic] from San Jose daily to San Francisco to work in his office. He (the W2 employee) cannot use his travel, or commute expenses from home to work, neither can you.’.”); USER162 (“Well… be careful on that one. Commuting miles are never deductible. If its from first job to second job, then yes. There is a fancy chart in the IRS about it.”).
 The following exchange reflects this confusion:
USER152: “Dead miles don’t count. That’s fraud. Find me where you can legally claim dead miles.”
USER125: “If you drive around looking for fares that’s a valid business purpose. Show me where it’s not.
USER152: “Show me where it is. You are not actively engaging in commerce unloaded.”
USER100: “No. Dead miles is called cost of doing business and /or commuting to work. . .”
See also USER9 (“You also you [sic] can only deduct mileage en route to a pick up, on a run, or returning to the start of a run. You cannot deduct mileage driving around while on line.”).
 For example, some participants engaged in such reasoning by analogy on one thread:
USER135: “If you drive your car to meet with a client for sales or to go to a business meeting your miles are deductible. So what’s the difference in driving to pick up a rider?”
USER152: “In the case you described the business owner has a home office that is ‘brick & mortar.’ You don’t. It’s the same as if you had a paper route. You can count your miles on your route and that’s it. If you have questions like this, you should consult the IRS publications. The only way you can claim dead miles is if you have a commercial reason to count them. The reason why you can’t count the miles to pick up a rider is because you don’t start the meter until they get in the vehicle . . .”
USER135: “I don’t see the difference between driving around hunting down fares and a salesman driving around looking for sales. If I am on the clock, burning gas in the process of my job, it’s a business expense. Plain and simple . . .”
 See, for example, the following exchange:
USER25: “The only mileage you can keep track of is the mileage where you’re making money. You can’t use mileage unless you’re actually in the middle of a fare…. I was shocked when my tax lady explained this.”
USER78: You need a new tax lady. This is absolutely false. You can deduct all the miles between when you leave to start Uber driving until you return home. If you make personal stops during that time, those should not be deducted, however.”
See also supra note 109.
 See, for example, the following exchange:
USER10: “I just spoke to a Accountant, you need to be able to prove you started at your house. He said I can count point A to B to A and so on, Everything but you milage before you [sic] first and after your last ride.”
USER156: “Yeah that is what I thought. Ill have a hard time documenting that, so Ill ask my guy to see what the [sic] thinks. In years past we have just come to an agreement with the auditor on things like this, so Ill keep logging all my miles.”
Following this exchange, another driver chimes in.
USER9: “Not according to my tax man. Any time you are not attached to a run (the point you accept the ping), or returning to the starting point of a run, it is considered commuting time and is not deductible.”
See also generally exchange between USER98 and USER8 regarding conflicting advice given by their respective accountants.
 See, e.g., USER176 (“Former CPA here – fyi, you can write off any business mileage, which includes driving from home to your area of work, as well as driving around looking for passengers. Uber will only tell you mileage while you have passengers, which is often a lot less than the mileage you can write off.”).
 See discussion infra note 252 and accompanying text.
 See, e.g., USER167_TO (“My strategy…was to stay online at my house until pings came in. I live very close to two popular bars in town, as well as relatively affluent residential area. After completing trips, I would look to return home and if I got a ping on the way from either of the bars or houses around I would take it. . . . I drove Friday and Saturday night, and was online on the driver app for about a total of 8.5 hours. My total time where I was earning money during trips was about 3:25 minutes. I would say the rest of the time was spent between driving get pickups or watching TV at my house waiting.”).
 See, e.g., USER87; USER3; USER151 (stating that they deducted “driver ‘uniform’”); USER138 (stating that they would deduct haircuts and clothes “to ‘keep me professional’”); USER40 (writing off clothes and internet).
 See, e.g., USER71 (“Haircuts and clothing are generally not deductible. Clothing can be if it is required specifically for the job – think welding helmet or fire retardant jacket.”).
 USER146; see also USER168 (deducted meals).
 USER172; see also the following exchange: USER90 asks, “Can I claim my daily coffee as a deduction”; USER70 and USER44 say yes; USER65 says no.
 See discussion infra Part V.C.
 But see the following exchange: USER149 states: “My understanding is that all mileage while ‘on the clock’ can be deducted, not just that while driving a passenger.” USER143 responds: “The first pickup, last drop-off, and being at a ‘busy area’ are definitely a ‘business location’. It’s a big ‘gray area’ if you can count miles before your first pickup/arrival at a ‘busy area’.”
 See, e.g., USER83 (“I’m not claiming to have the answer, but from what I’ve read, we . . . are supposed to use a schedule C to report this income. This apparently is also where we report our expenses (ex: mileage)”; USER82_TO (“I drove for Lyft for a couple of months last year. I was wondering what if anything I was allowed to write off. Gas, Service on my car, Miles.”).
 See, e.g., Answer by USER155; Answer by USER154. We counted 29 instances of formulaic language in TurboTax.
 Slemrod et al., supra note 49. That same study observed that Form 1099-K might incentivize taxpayers who had not previously filed Schedule C (Profit or Loss from Business) to file that schedule. Id.
 See infra Part VI.A.
 See, e.g., USER48 (“You can write off a work vehicle purchase in full the year you buy it, or you can deduct a partial amount each year via depreciation.”); USER6 (“but what if I were to deduct out of my income the cost of a vehicle? say I make $15,000 this year and buy the car I’m driving for $15,000. is that any different than buying gas and bottled water? It’s done for business purposes.”); USER120 (“I’m not a tax expert either, but I think it depends on how much you use the car for business versus personal use. If you use the car half and half, say, I imagine at best you could deduct half the cost as a business expense, not the full cost.”).
 See supra note 96 and accompanying text.
 See, e.g., USER20 (explaining standard mileage rate); but see USER165 (stating that standard mileage rate is unavailable to rideshare drivers); USER64 (“Can you explain the ‘standard mileage reduction method’? And how much of it is a tax write off? I’m so new to this, it’s scary what I don’t know. Thanks in advance for explaining like I’m 5.”).
 See, e.g., USER173_TO (“How do you guys handle your 1099 tax status? Totally new to this. No idea how it works, never been an independent contractor. I don’t really have a ‘tax guy,’ Im just a ‘go to Liberty taxes and get em done every year.’ Im in LA, if that makes any difference. Any stories/ideas/tips? Anyone ever just not even file?”); response by USER78 (“Not filing is a terrible idea. It’s pretty easy. Liberty, H&R Block, Turbotax, will all be able to handle the taxes without batting an eye. Make sure you track your expenses and take all the deductions you’re entitled to. Obviously consult a qualified professional, but generally the following are deductible:
*IRS mileage rate for all the miles you drove related to Uber, including mileage between passengers.”).
 See, e.g., USER77 (“If you choose to use itemized expenses in order to deduct those costs, you have to use it every year for the duration of the lease. You could also forgo all of that and just use the Standard Mileage Rate. But with a leased car, you have to use the SMR every year for the duration of your lease. You can’t switch to itemized in a later year like you can with a car you own.”); USER99 (“…if you start by itemizing you cannot switch to the standard $0.56 deduction the next year…”).
 USER15. Although the comment suggests that the poster is using the standard mileage rate as a benchmark for calculating profitability (which several posters seem to do), it also implies that the poster does not anticipate additional deductions for cash washes, and that such outlays are included in the $.56/mile.
 See, e.g., USER5 (“I would think that you could argue car washes are not part of standard mileage and therefore deductible.”); USER121_TO (deducted car washes and took standard mileage).
 USER23 (noting that if one gets audited, one can “create [one’s] own documentation”; stating that drivers should “estimate” and that “[t]he chance of being audited…is very slim.”).
 We counted 83 instances of the code “audit risk” in Reddit, and 160 instances in UberPeople.
 See, e.g., USER128 (“Just as a side note, make sure your logs include:
- Starting odometer reading for each trip
- Ending odometer reading for each trip
- Start and End addresses of each trip if applicable
- Total mileage driven for the trip
- Purpose of trip
Less information than that, and your logs might not be usable in an audit, I believe. IANAL [I am not a lawyer]”); USER162 (“Technically you could recreate a written mileage log. The uber app shows your first trip of the day. You can calculate the start location and determine the distance from when you first click ‘on’ which is probably at your home or office (if you have a regular job). You could reconstruct your mileage that way. But the better plan is to have a mileagein, mileageout log, with the beginning/ending odometer readings. I typically use one of these in excel, and have multiple columns in the middle to allocate the mileage between various categories (uber, workreimbursable, worknonreimbursable, commuting, charitable, etc).”)
 For example, consider the following exchange:
USER176: One other note the IRS doesn’t accept mileage that is tracked in a spreadsheet, so the spreadsheet wouldn’t hold up as proof if you were ever audited (meaning potential back taxes and penalties).
USER166: “how does the irc accept mileage then???”
USER176: “Their preferred method is actually a paper log (yes, very outdated). We have basically made an electronic version of this that tracks your odometer over time. They haven’t explicitly commented on if mileage tracking apps are sufficient proof or not.”
USER77: “Thats exactly why my excel document contains a printable weekly log to be kept in your car and filled out daily….You would then transfer that data into the Mileage & Expense Log on a weekly basis, so the other worksheets in the document can pull the data needed for other calculations (so you dont have to do it yourself). You would also keep the handwritten log as backup in a safe place. however, the IRS directly states that “You must generally prepare a written record for it to be considered adequate. This is because written evidence is more reliable than oral evidence alone. However, if you prepare a record on a computer, it is considered an adequate record.”
See also USER139 (“At the end of the year you only need the business miles. If you need to defend the return, the daily log is priceless. You can also use My Tracks or whatever Google calls it, this will log an entire route and default to saving it by date and time. But I think writing it down is still the best way, so far.”).
 See discussion supra Part II.A.2.
 See, e.g., thread entitled “Basically…does Uber screw us?”; USER17; USER114.
 See, e.g., USER78.
 See, e.g., USER77. USER78 and USER77 get into an argument about whether the fact that the Form 1099-K lists gross amounts matters in the end. USER78 gets it. USER77 does not.
 See, e.g., posts entitled “Anyone else received a screwed up tax summary?”; “Check your 1099s”; “1099-K and 1099-Misc WTF???”; “1099-K has incorrect numbers”; “Check your 1099K **carefully** – they screwed mine up”; “You will get 2 forms! 1099MISC and 1099K”; “Here, From The IRS, Is The Link Detailing What a 1099K Is, And What To Do With It”; “Text post about my 1099 form and IRS”; “1099K and Taxes.”
 See, e.g., USER153 (“How to read your Uber 1099”); USER176 (“We wrote a blog post on how to read your Uber 1099 – check it out here! http://tryzen99.com/blog_posts/readuber1099”); USER176: post entitled “How to read your Uber 1099” and link to blog post). There are numerous instances of commenters explaining that drivers have to report the gross Form 1099-K amounts and then deduct fees. See, e.g., USER78 (“Uber now reports all your earnings except for referral bonuses and a few random things on the 1099K. You have to report that income and then deduct the various Uber fees from that.”)
 The literature on Form 1099-K reporting focuses on how increased expense taking may offset revenue gains but has not discussed the possibility of overreporting. Slemrod et al., supra note 49.
 TurboTax Lyft P5, TurboTax Lyft P6, TurboTax Lyft P11, TurboTax Lyft P12, TurboTax Lyft P16, TurboTax Lyft P17, TurboTax Lyft P18, TurboTax Lyft P19, TurboTax Lyft P21, TurboTax Lyft P23, TurboTax Lyft P25, TurboTax Lyft P27, TurboTax Lyft P30,
 See, e.g., posts entitled “1099Ks Uber vs. Lyft” and “Basically…does Uber screw us?”
 USER7; USER57.
 We discuss our findings regarding worker classification issues in other work.
 See sources cited infra notes 187-190 and accompanying text.
 Of the 13 comments, the handle USER164 was responsible for 7.
 USER126 (“But…I don’t want to be an uber employee”).
 USER164 (“…you can also be held liable for back taxes”; “[t]he best part of this Uber thing is not having a boss!”); USER50 (“…employers will always pay as little as they can legally get away with. The difference is they can now dictate your work schedule and behavior on the clock”); USER31 (“there are no managers we dont’ report to anyone”).
 USER130. The other poster pointed out the ability to unionize as an upside. USER109.
 See sources cited infra notes 187-190 and accompanying text.
 See, e.g., USER51 (“Your best strategy is to consult with a tax accountant a few months before your taxes are due and track and keep ALL your receipts until that time.”); USER41 (“I would, highly, and I MEAN HIGHLY suggest you go with a real tax professional while working for uber.”); USER170 (“Get a reputable tax person and do exactly what they tell you.”); USER46 (“I have an accountant that will be reviewing my taxes before filing….”); USER5 (“…I would advise getting a tax professional….”); USER158 (“…Talk to your tax person,…”); USER77 (“Consult a professional.”); USER86 (“Get a CPA. Use reddit advice for asking your CPA questions, but don’t take tax advice from people so bad at math they drive for Uber.”); USER140 (“It is very much worth your time and money to speak to a Certified Public Accountant.”).
 USER148 (“…absolute bullshit how much they charge for someone to input numbers.”); USER62 (“HR Block and other similar 3rd party tax preparers are getting the best $ from this paid $200 for a $111 return”); USER136 (“How much does H&R Block charge to do your taxes?”); USER47_TO (“Ok so I am looking for what name, address, and EIN that appears on the 1099’s Uber sends out. I know that you don’t need it to file but you do if you don’t want your account/tax software to charge you a higher price.”).
 See USER53 (“ill contact my accountant and see his thoughts. I work in a business with many types of write downs and have to go to a special accountant because of it. A few years back I switched accountants and my current guy told me he could have saved me an extra $3000 if i had written this and that off. Saving thousands is as simple as being with a guy who has the knowhow versus being at the wrong guy. Maybe this works the same.”); USER117 (“Of course im not tax person and even the tax person I went to didn’t seem that bright”).
 USER51, USER59, USER121, USER78, USER79_TO; USER144; USER134.
 TurboTax Uber P20, TurboTax Uber P42, TurboTax Uber P43, TurboTax Uber P44, TurboTax Uber P45, TurboTax Uber P47, TurboTax Uber P53, TurboTax Uber P60, TurboTax Uber P61, TurboTax Uber P65, TurboTax Uber P67, TurboTax Uber P76.
 USER69 (“I just worked for Uber as an independent contractor so I don’t consider this running a business…,”); USER96 (“I work for uber…. So i dont have my own business so i dont need the upgrade to business but it is making me i dont know Why”). See also TurboTax Uber P6, TurboTax Uber P42, TurboTax Lyft P4, TurboTax Lyft P9, TurboTax Lyft P18.
 See generally Laura Northrup, Changes to TurboTax Lead to Consumer Revolt, Opportunity for Competitors, Consumerist (Jan. 23, 2015), http://consumerist.com/2015/01/23/changes-to-turbotax-lead-to-consumer-revolt-opportunity-for-competitors/ [http://perma.cc/C4FW-P7JK]; Janet Novack, Intuit Cries Uncle, Will Reverse TurboTax Deluxe Changes, Forbes (Jan. 29, 2015), http://www.forbes.com/sites/janetnovack/2015/01/22/intuit-offers-25-refund-to-turbotax-deluxe-users-hurt-by-software-changes/ [http://perma.cc/T9NV-GAJF] (Deluxe users could fill out Schedule C as a form, but the Home and Business upgrade was required to file it as part of the TurboTax package).
 Novack, supra note 167.
 See, e.g., USER106 (“I once made a spreadsheet to figure in depreciation, insurance, gas, oil changes, tires, etc…and it costs me like $.20 a mile to drive.”); USER118 (considering price of car, gas, tires, insurance, brakes); USER110 (“You know that Uber never considers driver expenses such as fuel, maintenance or depreciation”); USER58 (considering gas, wear and tear, and taxes).
 See, e.g., USER27 (“This is emberassing, but I’ve never really had a job, hence income. What does it mean to ‘write off taxes’? Can someone ELI5 [explain like I’m five]? I’ve no experiences in taxes either. Considering becoming a driver.”); USER116_TO (“I just recently became a driver and didn’t read over the 1099 tax form because I needed to make money quickly. How do I know if I have to pay into taxes or not?”); USER21_TO (“…When does Uber pay the driver? Are taxes are taken out of the payment to the driver?”); USER104_TO (“Do you drivers usually owe (pay taxes) to the IRS, or do you get money back? (Tax refund).”).
 See, for example, the following exchange:
USER175_TO: “So I’m trying to figure out how this 1099 thing works with Uber. To my understanding, if I were to make $20,000 and had more than 200 transactions, then I would be sent a 1099-K. If Uber paid me $600 or more directly, then I would get sent a 1099-MISC. Does this sound correct? Does this mean if I make less than $20,000 then I don’t need to file? Also, what direct Uber payments count towards the 1099-MISC? Does anyone in CA have experience dealing with 1099’s from Uber?”
USER40: “of course you have to file taxes if you make less than $20k!! lol talk to a tax account in your area.. dont believe what you read on reddit about taxes.. seek a professional.. i do know you should mail 20% of your gross every 3 months using the 1040EZ form and in NJ 5% to the state but thats it..”
USER161: “Correction especially for part-timer Uber drivers, you only have to pay the estimated quarterly tax if you believe that you will owe more than $1000 in taxes at the end of the year. With all of the mileage deductions, car washes, and rate cuts, unless you are Ubering full-time you might not owe more than $1000. Remember it’s not $1000 in Taxable Income, it’s OWING $1000 or more at the end of the year.”
USER40: “yea but most of us are gonna owe more than $1000.”
See also USER101_TO (“I just got approved to drive a week ago, so I have yet to figure out the tax implications of my shiny, new parttime employment as an Uber driver….I know that fulltime Uber Drivers are typically classified as independent contractors and are required to pay quarterly estimated taxes once they pass a certain profit threshold however, is that still the case for someone who has a fulltime, 40+ hour job and is just doing Uber on the side for extra cash?…I’m currently planning to set aside 30% from every Uber paycheck in a separate account so I can settle up at tax time, but as for whether I should be paying quarterly estimated taxes or whether I should just be reporting this as additional income and paying what I owe come April 15th I’m at a loss. I also have no idea how it will affect my filing for my regular fulltime job. (Why must the tax code be so stupidly complicated?)”); USER94 (“Guessing a majority of Uber drivers are going to have an extremely unhappy tax season. Guys driving a new Prius through uber finance and driving uber full time, they’ve likely been living off of money that should have been withheld for taxes. I bet we’ll see a large wave of driver defections in the next few months because of this.”).
 USER74 (“My thought is to just toss 20% in a savings account until I file taxes next year, but will this be enough?”); USER77 (“I am currently preparing to put away about 30% of my total uber income for taxes.”); USER92 (“I saved 25% of my uber payouts anticipating taxes.”).
 See, e.g., USER55 (“You only need to pay taxes if you make a profit. With Uber’s wonderful rates I don’t need to worry about that one”); USER159 (“If you are paying taxes on Uber income, you’re doing it wrong.”).
 See David Gamage & Darien Shanske, Three Essays on Tax Salience: Market Salience and Political Salience, 65 Tax L. Rev. 19 (2011); David Gamage, On the Future of Tax Salience Scholarship: Operative Mechanisms and Limiting Factors, 41 Fla. St. U. L. Rev. 173 (2013); Brian Galle, Hidden Taxes, 87 Wash. U. L. Rev. 59 (2009); Deborah H. Schenk, Exploring the Salience Bias in Designing Taxes, 28 Yale J. on Reg. 253 (2011); Jacob Goldin, Sales Tax Not Included: Designing Commodity Taxes for Inattentive Consumers, 122 Yale L.J. 258 (2012); Lilian V. Faulhaber, The Hidden Limits of the Charitable Deduction: An Introduction to Hypersalience, 92 B.U. L. Rev. 1307 (2012); Andrew T. Hayashi, The Legal Salience of Taxation, 81 U. Chi. L. Rev. 1443 (2014); Susie Cleary Morse, Using Salience and Influence to Narrow the Tax Gap, 40 Loy. U. Chi. L.J. 483 (2009).
 This terminology comes from Liebman and Zechkauser. Jeffrey B. Liebman & Richard J. Zeckhauser, Schmeduling (2004) (unpublished manuscript) available at http://www.hks.harvard.edu/jeffreyliebman/schmeduling.pdf [http://perma.cc/4NVJ-3HT2].
 Id. (“Spotlighting occurs when consumers identify and respond to immediate or local prices, and ignore the full schedule, even though future prices will be affected by current consumption.”); Gamage & Shanske, supra note 175.
 Gamage & Shanske, supra note 175, at 27.
 Spotlighting behavior has been empirically demonstrated primarily in the sales tax, excise tax, personal property tax, toll charge, and tax incentive contexts, Raj Chetty, Adam Looney & Kory Kroft, Salience and Taxation: Theory and Evidence, 99 Am. Econ. Rev. 1145, 1165 (2009) (showing grocery shoppers were less likely to purchase when tax information was posted than when only pre-tax prices were posted, despite having accurate knowledge of sales taxes. Also demonstrating that tax increases included in posted prices reduce alcohol sales more than taxes applied at the checkout); Richard Ott & David Andrus, The Effect of Personal Property Taxes on Consumer Vehicle-Purchasing Decisions: A Partitioned Price/Mental Accounting Theory Analysis, 28 Pub. Fin. Rev. 134 (2000) (“Although respondents opined that [vehicle personal property taxes] were too high, they had little effect on their decision to purchase a vehicle.”); Kelly Sims Gallagher & Erich Muehlegger, Giving Green to Get Green? Incentives and Consumer Adoption of Hybrid Vehicle Technology, 61 J. Envtl. Econ. & Mgmt. 1, 9 (2011) (finding that “[c]onsumer response is greater to a sales tax exemption, which is immediate and automatic at the time of purchase, than to an income tax credit, for which a consumer must understand, apply for, and eventually collect as part of their tax return”); Amy Finkelstein, EZ-Tax: Tax Salience and Tax Rates, 124 Q.J. Econ. 969, 970 (2009) (finding that when using electronic rather than hand toll collection, “driving becomes less elastic with respect to the toll and toll setting becomes less sensitive to the electoral calendar”). See also Tomer Blumkin, Bradley J. Ruffle & Yosef Ganun, Are Income and Consumption Taxes Ever Really Equivalent? Evidence from a Real-Effort Experiment with Real Goods, 56 Eur. Econ. Rev. 56 (2012) (finding that “the temporal separation between an individual’s labor market allocation and subsequent consumption decisions leads individuals to work longer when faced with a consumption tax than with an equivalent wage tax”); David Gamage, Andrew T. Hayashi & Brent K. Nakamura, Experimental Evidence of Tax Salience and the Labor-Leisure Decision: Anchoring, Tax Aversion, or Complexity?, 41 Pub. Fin. Rev. 203 (2012) (finding that “[s]ubjects are most willing to work when their net wage is transparent. Any additional complexity in the wage description in the form of either a low base wage plus a surcharge or a high base wage less a tax, decreases work participation”).
 Gamage & Shanske, supra note 175, at 29 (“If workers make job choices based primarily on posted pretax salary information, rather than on their aggregate post-tax salaries, then even the income tax may have low-market-salience”); Jacob Nussim, To Confuse and Protect: Taxes and Consumer Protection, 1 Colum. J. Tax L. 218, 253 (2010) (noting workers may be misled by gross salaries and may choose to oversupply labor, but arguing that this may be an optimal outcome).
 Oei & Ring, supra note 5; Barry & Caron, supra note 5.
 See, e.g., sources cited supra note 173.
 See generally sources cited supra notes 175 and 179 for a discussion of ironing.
 See supra note 162.
 See generally supra Part V.A.
 See discussion supra Part V.A.5.
 See, e.g., USER127 (“So now what? Uber and Lyft are going to be hit with a shitload more expenses, and it’s coming straight out of your fares, because they’re certainly not going to raise the rates.”); USER12 (“Personally am not a fan of unions, would not join or pay dues, I am an independent contractor.”); USER45 (“We’re independent contractors running our own businesses. We’re management, not labor.”); USER88 (“Personally, I have no interest in being an Uber employee”); USER117, USER169 (discussion of possible effects on Uber business model if drivers are held to be employees). One commenter noted, “I hate that we have a distinction, which is what’s driving my disdain in the first place.” USER127.
 USER55 (“I am not sure “siding with us” is the right term. Yes the government want’s a bigger chunk, I am not sure that will mean a bigger chunk for us as well. You can not force an evil corporation to treat you better.”); see also USER13, USER177, USER109, and USER159.
 USER150 (“Uber is exploiting a one sided coercive contract that favors Uber over the drivers”); USER80 (“…freelance workers would end up taking the biggest hit”).
 As an example of a poster not appreciating the impact of vehicle wear and tear, see USER35_TO (“I just got my 1099k and for the 4 months I was driving I made $3200. I haven’t added up my receipts for gas yet but I’m figuring around 500 or so spent. When I first started driving my car appraised for $18000. I later sold my car for $16000. Because of the mileage. They wouldn’t give me the same amount as the original offer. So all said and done after taxes I’ll actually end up loosing money driving for Uber.”). In response to comments from other posters who questioned why this forum participant was surprised that driving would likely decrease the value of the car, USER35 clarified: “I knew it would but I didn’t expect a depreciation of $2000. Once I get my taxes in line I’ll have a final number. I’d be happy coming out even. I needed the money when I was driving. It helped out a lot when I needed it but not worth it to me anymore. I got a new job making significantly more money.”); see also USER54 (“There’s no doubt that drivers have higher costs than your typical salaried 95’er, but that is the price we pay for having much greater freedom than the other guy. . .. On top of that, it’s not exactly equal because I didn’t take into account the much more frequent oil changes that I perform than your average 95’er, and car maintenance in general, but all of this goes back to the original question that I posed Is the flexibility important enough for you to take on these expenses?”).
 See, e.g., USER1 (“That’s $160 after commission. So assuming that’s 200 miles and operating cost is .57/mile as IRS deduction states, that’s $114 operating cost, and you’d be taxed on $46 (difference in revenue operating costs) which is $11.5. So net after commission, operating costs, and tax your take home is $34.5 for 10 hours of work.”); USER18_TO (“The IRS deduction rate of $0.575/mile for 2015 is pretty accurate in terms of the actual cost per mile to drive a vehicle after depreciation, gas, oil, maintenance, and other expenses.”); USER93 (“The standard mileage deduction (which I used ALL of the mileage that I clocked when the app was ON and ready for ride requests) worked in my favor. It showed that driving for Uber was a money loser for me.”).
 See sources cited supra note 192.
 See, e.g., USER42 (“The IRS mileage deduction isn’t ‘operating costs.’ It’s solely for the depreciation of the value of the vehicle due to mileage driven, and it’s only relevant at tax time for purposes of ‘reducing’ your tax bill. It certainly doesn’t cost me (or anyone else) $114 to operate a car for 200 miles.”); USER157 (“be careful. Many people in here will say they made $10 for a 10 mile ride, then subtract gas, then subtract 57.5 cents per mile, and tell you they made $2 an hour. They think the 57.5 cents is their ‘cost to operate their car’ because the ‘IRS says so.’”); USER85 (“I run my vehicle for way less than .56.”).
 USER124, USER56, USER94.
 In both Reddit and Uberpeople, the code “Profitability: Per-Mile vs. Per Hour” was among the most common (361 occurrences in Reddit; 359 in Uberpeople).
 The fact that vehicle wear and tear may ultimately be a bigger cost in ridesharing may be less salient to drivers because gas purchases constitute a daily or weekly purchase.
 See, e.g., USER142 (“since uberx decided to drop prices by 70% last year drivers are making less than minimum wage. when you work for walmart you are guaranteed you will make 9 dollars for every hour you showed up. with uberx sometimes you can come back home empty handed.”); USER99 (“…Uber is now the walmart of rideshare/transportation, they do not care if the drivers make livable wages, if you want a better job, look somewhere else.”); USER67 (“A Walmart checker may gross $2,000 an hour at their register, but what do they get to take home each hour?”). Of course, such comments do not reflect the greater flexibility in setting hours for Uber drivers than for substitute low-wage employment. See, e.g., USER133 (“I am not obligated to work Uber which is the bonus. Doing this and my full time job I work 6 days a week. It’s tiring and fortunately I can drop one whenever I feel like it.”); USER99 (“I would rather drive Uber than work behind the counter of a McDonalds”).
 Another choice discussed was whether to drive for UberX or UberBlack, or to do both.
 See, e.g., USER23 (“Uber’s meant to be a parttime, supplemental income gig. making it a full time, main hustle is where I see people get bitter about it.”).
 See, e.g., USER11 (“I like being able to pick and choose the times I drive. I only drive during surges now. i work like 8 hours a week usually making 200250. I tried driving 8 hours straight one day and I was do [sic] bored out of my mind. I like to think that I can just stop after one or 2 good rides.”); USER163 (“There’s just not enough weekday daytime business in Richmond to make Uber a viable fulltime gig.”).
 See, e.g., USER68 (“My advice: you work full time at this, you burn out. Everyone who does this part time enjoys it. Those who don’t…”); USER169 (“Also, in order to really make it work full time, you’d need to work a crazy schedule. It just doesn’t make sense to drive between 9 and 5 when everyone else is at work so you’d have to do a few hours in the morning, go home, come back out, do several more hours in the evening, then work every weekend. I don’t know about you but that’s not something I’d be willing to do.”)
 See sources cited supra note 191; see also USER73 (“I suppose I question as to whether or not you are really accounting for the wear and tear on your car driving UBER full time.”); USER174 (“It is very easy to become resentful of Uber and the Pax when its a f/t gig…. You will drive your vehicle into the ground, basically stealing its value to put in your bank…”).
 See, e.g., USER169 (“IMO, Uber isn’t reliable enough to turn it into a career. It’s great for cash on the side or to help you out if you’re in between things but, as we’ve seen in most parts of the country, rates can drop at a moment’s notice whenever Uber feels like it.”); Uberpeople P470-476 (thread containing various posts about rate cuts).
 See, e.g., USER42 (“I agree, and the fact that I have regular work/don’t need to worry about setting aside 35% for taxes is why it still pays for me.”); USER 68 (notes that you can adjust job withholding to avoid paying quarterly taxes on Uber income); USER36 (“If you are parttime, I assume you have a W2 job. Instead of doing estimated tax payments, you can adjust the W2 tax withholding to cover both jobs.”).
 Although forum participants did not explicitly state that these benefits were untaxed compensation, the point was implicit. See, e.g., USER169 (“It really depends on what your current job is. If you’re making a living wage, have job security, benefits, time off, vacations, etc… Then stick with it even if you think you could make a little more in the short term with Uber.”); USER137 (“Because of the fact that just about any entry level, fulltime job pays better than Uber and has actual benefits, no person who is qualified for and able to acquire such a job would drive for Uber full time.”).
 See generally George Borjas, Labor Economics (7th ed. 2016).
 See, e.g., USER24 (“As a result I have taken a view to only drive for pure sport and entertainment…and as a loan until I have to file for taxes.”); USER81 (“…the hourly rate isn’t bad for a part time job. I also get to know my city better and I can meet and talk with all kinds of interesting people while on Uber time.”); USER2 (“Uber is like gambling. Some people do it for a full time job but most just do it for fun.”); USER37 (“…regardless I’m not doing it entirely for the cash. I don’t get to drive as part of my daily routine, so 58 hours a week behind the wheel is kind of fun. Plus I’m going to enjoy getting to know other parts of the city. And finally, I hope to meet some interesting people. I think it’s a win/win.”); USER160 (“Same. I do it for ‘fun’. I enjoy the people I have given rides to, the stories I have and the rush of picking up a fare. It gives me a few extra bucks some weeks with the flexibility to do a ton of driving or zero. Would I ever do Uber for a full time job? Heck no.”); USER169 (“I enjoy driving and meeting people and the social aspect of Uber as well, but based on what I’ve seen around this sub, I suspect I’d feel differently if I put 40+ hours in every week instead of ~10.”).
 The term “beer money” occurred frequently as shorthand for driving for extra cash. See, e.g., USER108 (“Because most UberX Drivers are like myself. People who just do this part time while family is asleep to make some extra beer money.”); USER137 (“ . . .Uber knows this and exploits the fact that its parttime drivers are either desperate for money or say ‘it’s for beer money’ and don’t bother to do the calculations to realize they are making below minimum wage.”); see also UBER42 (“For me, it’s exactly what it was supposed to be when uberX first came out extra pocket money, not a career.”)
 See discussion supra Part V.B.1.a.
 See, e.g., USER29 (“how do I add mileage to my return? I am an Uber driver and need to add mileage to my return”); USER33 (“Does Uber count as public transportation for tax purposes?”).
 USER151 (attempts to deduct uniform); USER138 (will deduct haircuts and clothes); USER28 and USER71 correct USER151 and USER138, saying that these items are not deductible; USER40 (says he will deduct clothes, internet, and other items); USER107 (sarcastically questions these deductions by USER40).
 USER168 says they are going to deduct meals; USER78 says “You can’t generally deduct meals”; USER4 says “you must be a really shitty uber driver if youre deducting meals from your tax. FRAUD”. See also the following exchange: USER3 claims that he deducts one meal a day; USER78 informs him that “Meals in this circumstance are absolutely not deductible.” USER86 also says meals are not deductible.
 USER6 (thinks he does not have to pay taxes, other forum participants correct him).
 USER7 (asks whether income not reported on Form 1099 needs to be “claim[ed]” and USER66 suggests that “[i]f [Lyft is] not reporting it to the IRS then I’m not going to report it”; USER153 (tryzen99 founder) alerts them to compliance obligation and audit risk).
 See also the following exchange:
USER113; “If you have another vehicle in the family available for personal use (I.e. for me my wife has a truck) then you can claim every mile you drive your user vehicle as business whether it’s actually for business or not. You may want to fabricate a mileage log to back that up though just in case”
USER113: “Yes you can claim 100% business usage on a vehicle if is not your only vehicle …why would that be false?”
USER162: “That is true but you said: You can claim every mile you drive your user vehicle as business whether it’s actually for business or not. The ‘or not’ makes it false.”
USER162: “…And fabricating a mileage log is fraud. Play by the rules. You would be surprised what information comes out when IRS CI investigates a TP. Crazy the stuff people will say on an open forum.”
 See, e.g., Uberpeople P11-P13 (several misstatements about whether coffee, car washes, and meals are deductible go uncorrected, or, even if corrected, are overridden by other erroneous posts).
 On one thread, for example, USER153 and USER98 warn about commute miles not being deductible, but USER85, USER61, USER92, USER8, and others persist in claiming that all miles driven with the Uber app on are deductible and that there is no commute.
 See, e.g., USER162; USER105; USER51; USER145.
 See, e.g., USER145; USER91; USER60; USER78, USER38.
 See, e.g., USER77_TO; USER89; USER97; USER72; USER49; USER34.
 USER161. This statement and the next are oversimplifications of the hobby loss rule. I.R.C. § 183.
 USER23. This comment does not comply with the substantiation requirements.
 USER165. This statement is wrong. Rideshare drivers can use standard mileage.
 USER3. Meals are not deductible.
 USER19. This is obviously not the correct way to track mileage. See, e.g., Car and Truck Expense Deduction Reminders, Internal Revenue Service, http://www.irs.gov/uac/car-and-truck-expense-deduction-reminders [http://perma.cc/3TNV-XG6F] (last visited Dec. 8, 2016) (taxpayer should maintain a daily mileage log).
 USER44. Meals are not deductible. See, e.g., I.R.S. Publication 463, Travel, Entertainment, Gifts, and Car Expenses (2015) at 5 (meals are deductible if taxpayer travels overnight for work, or if the travel is long enough to require “sleep or rest”).
 See, e.g., USER120 (“I’m not a tax expert either …. But if you get audited don’t look at me. You’re better off Googling stuff or asking a tax adviser.”); USER68 (“As I said, this is what I’ve gleaned, but I’m not a tax professional. Feel free to correct me if I’m wrong”); USER86 (“Pro tip: Don’t take advice from this sub”); USER141 (“I start [tracking miles] from the time I go online to accept calls. If I get no action, and need to move closer to a busy area, I include those miles too. However, I’m not a tax professional.”). See also the signature block of USER75, a well-known forum member. (“I am a geologist; not a CPA, attorney, or anything else that means my work should be taken as the gospel. I provide information/advice to the best of my knowledge from my personal experience driving for UberX & Lyft and business knowledge”); USER107 (“Asking for tax advise on this (or any) forum is a recipe for disaster and an unpleasant audit. Plenty of bad advice will be given. It is best to ask a tax professional (hint: They’ll just say to take the IRS standard $0.56 per mile)”).
 See sources cited supra note 232.
 See, e.g., USER153; USER131; USER95; USER122; USER176.
 See, e.g., USER77; USER105; USER51; USER111.
 This poster used two separate usernames, one in each forum.
 As of August 2015, tryzen99.com shut down its business. Biz Carson, There’s A Big Shift Going on in Startups, and this Company Just Got Caught in the Middle, Bus. Insider (Aug. 12, 2015, 6:06 PM) http://www.businessinsider.com/why-zen99-closed-its-doors-2015-8 [http://perma.cc/5WN8-S2GK].
Zen99, http://web.archive.org/web/20151022200742/http://tryzen99.com/1099-tax-guides/ridesharing [http://perma.cc/7Z8E-SR3U] (last visited Dec. 8, 2016) (“You can … write off any snacks or amenities purchased for clients (e.g., water).
 “[Y]ou can write off any business mileage, which includes driving from home to your area of work, as well as driving around looking for passengers.”; “Since you are an independent contractor, you can write off any business related mileage. This includes driving from your home to area of work, or driving around waiting for passengers.”.
 See supra note 98 and accompanying text.
 This user said in response to critiques of prior posts, “Most of those commuting rules are for employees who travel to the same office. If they travel to a client office, they can write that off as business mileage. Technically, since you don’t have an office, every time you drive to the area is for ‘client’ business. The IRS hasn’t specifically commented on ridesharing, so some accountants will say ‘yes it’s fine’ and some will say ‘no it’s not’. Most ridesharers write it off though.”.
 Zen99, http://web.archive.org/web/20151022200742/http://tryzen99.com/1099-tax-guides/ridesharing [http://perma.cc/RFR8-UAG4] (last visited Dec. 8, 2016).
 It is possible that this position might have evolved over time, but we cannot tell the exact dates of some of the Reddit posts.
 We coded 90 instances on Reddit and 86 instances on Uberpeople of “marketing posts.”
 See, e.g., Reddit P14 (Hurdlr rideshare guide); P62 (app for freelancers), P72 (expense tracking excel file), P73 (feedback on whether to write ridesharing guide), P76 (sherpashare tax guide), P79 (feedback on starting tax firm focused on ridesharing), P83 (marketing for Shared Economy CPA), P94 (kickstarter project for automated trip logs app), P157 (seeking feedback on mobile expense tracking, etc. app for drivers); P255 (advertising app to combine fuel with mileage); P265 (sherpashare time/mileage/fare comparison feature between Uber, Lyft, Sidecar).
 USER162. IRS Circular 230, 31 C.F.R. Subtit. A. Part 10, http://www.irs.gov/pub/irs-pdf/pcir230.pdf [http://perma.cc/3NMY-D346].
 See generally Part V.A.
 Slemrod, supra note 49; see also discussion supra Part V.A.
 App providers such as MileIQ have claimed that their GPS tracking application will be acceptable upon audit. Stephen Fishman, Mileage Log Template: 4 IRS Requirements for Business Miles, MileIQ: IQ Blog, http://www.mileiq.com/blog/mileage-log/ [http://perma.cc/4M9L-Y2LK]; see also James Alm & Jay A. Soled, The Internal Revenue Code and Automobiles A Study of Taxpayer Noncompliance, 14 Fla. Tax Rev. 419 (2013) (arguing that GPS technology could be used effectively in ensuring accuracy of automobile deductions in general).
 See discussion infra Part II.A.2; see also Oei & Ring, supra note 5 (proposing some of the same suggestions).
 See, e.g., Australian Taxation Office, The Sharing Economy and Tax, http://www.ato.gov.au/Business/GST/In-detail/Managing-GST-in-your-business/General-guides/The-sharing-economy-and-tax/ [http://perma.cc/87PZ-CBT5]; see also HM Government Response to Independent Review of the Sharing Economy (Mar. 2015), 7-8 http://www.gov.uk/government/uploads/system/uploads/attachment_data/file/414111/bis-15-172-government-response-to-the-independent-review-of-the-sharing-economy.pdf [http://perma.cc/WU9N-X55K] (expressing UK HMRC’s intent to “mak[e] it easier for people participating in the sharing economy to understand their tax obligations and report their income to HMRC”); see also HM Revenue & Customs, Tackling the Hidden Economy: Extension of Data-Gathering Powers (July 22, 2015), http://www.gov.uk/government/uploads/system/uploads/attachment_data/file/447718/Tackling_the_hidden_economy_-_Extension_of_data-gathering_powers.pdf [http://perma.cc/W4QN-EKXS]. Denmark’s taxing authority is also looking into how to more effectively tax the sharing economy. See Philip Tees, SKAT looking into tax changes for Airbnb, Uber and other sharing economy services, CPH Post (Jan. 24, 2016), http://cphpost.dk/news/business/skat-looking-into-tax-changes-for-airbnb-uber-and-other-sharing-economy-services.html [http://perma.cc/2WZB-NM3X].
 Internal Revenue Service, supra note 14.
 This assumes that the Form 1099-K position taken by the ridesharing businesses is correct.
 See supra note 146 and accompanying text.
 See sources cited supra note 148.
 For an example of the 2014 Uber Tax Summary, see http://ridesharedashboard.com/2015/01/30/uber-taxes-2015-tax-summary-deducting-fees/ [http://perma.cc/UH4X-E3NP]. On the Tax Summary, the Uber fee and device subscription fee are not set out in the same column as the gross Form 1099-K amount. Uber merely states in parentheses and in a footnote in its Tax Summary that the Uber fee was included in the gross amount.
 A related literature, considering how compliance with the law can be a function of imitation and how that imitation can be an inaccurate rendition of the original actor’s compliance, could provide additional insight into crafting taxpayer guidance and education. See generally Bert I. Huang, Shallow Signals, 126 Harv. L. Rev. 2227 (2013).
 See generally IRS Social Media, IRS (Aug. 25, 2016), http://www.irs.gov/uac/IRS-New-Media-1 [http://perma.cc/E822-6UUQ]; Internal Revenue Service, Tumblr, http://internalrevenueservice.tumblr.com/ [http://perma.cc/XA2H-8EHF]; IRS (@IRStaxpros), Twitter, http://twitter.com/irstaxpros [http://perma.cc/SGW9-5ADF]; IRS (@IRS), Facebook, http://twitter.com/IRSnews [http://perma.cc/KAG3-BQ96]; http://www.facebook.com/IRS/timeline [http://perma.cc/R5YW-MU5G]. Social media interventions have been suggested in other jurisdictions as well. See sources cited supra note 260.
 See infra Part VI.C.
 See discussion supra Part V.B.1.a.
 This result may vary by state. See, e.g., Cal. Lab. Code § 2802 (West 2016) (requiring an employer to indemnify employee job expenses).
 See sources cited supra note 267.
 See sources cited supra note 260.
 The Consumer Financial Protection Bureau runs this type of Questions and Answer service. AskCFPB, Consumer Financial Protection Bureau, http://www.consumerfinance.gov/askcfpb/.