Flying solo: data show lawyers can earn a decent living on their own

Does it make sense to become a solo practitioner?

The law school critics predict doom if you make that choice. But to reach that conclusion they rely on data that is problematic and thus likely to mislead anyone who relies on it alone. But other, perhaps better, data – generated by the U.S. Census Bureau – shows that solo practitioners likely are doing reasonably well.

Let’s walk through the numbers.

First, the law school critics point to data from the IRS (summarized in a book by Professor Ben Barton of the University of Tennessee called Glass Half Full (“GHF”)) suggesting that in today’s (2012) dollars the income of solo practitioners has fallen dramatically relative to what lawyers earned in the late 1980s.

To reach this conclusion the critics select carefully from Barton’s IRS data. They start at the peak of lawyer earnings in that earlier period – 1988 – when the economy was in the midst of a bubble led by real estate and leveraged buyouts. No doubt, then, there was plenty of work for lawyers. In 2012 dollars lawyers in solo practice apparently (I say “apparently” because of problems with the data discussed below) earned (on average in means) about 70K then (in 2012 dollars) and in 2012 they apparently earned 49K. That would be a pretty steep drop (if it is accurate). One way of putting would be to say that lawyers in solo practice now only earn about 70% of what they once earned measured in buying power.

But hold on.

The very year after that 1988 high point the bubble burst and in 1989 solo’s were only earning 61K (again, in 2012 dollars). And if you go back a few years from 2012 to the near peak of the recent credit bubble (say, 2005) solo’s were earning 57K. In other words they had, over a 26 year period, lost only a tiny amount in buying power. Of course, I am cherry picking, too, in order to highlight the problem that can arise with this particular data set.

Since the critics rely on this data to suggest that there has been a massive structural shift against the interests of solo practitioners (using words like “collapsing economics” to describe the plight of solos) and lawyers generally it’s important to see that, in fact, that’s not really true.

Take another example. Go back a few years earlier in the data (by the way, thanks to Professor Barton for sharing his full data set with me – he did the hard work – and of course the conclusions I draw from the data are purely my own). In 2012 dollars, solo practitioners were earning 41K in 1982. In 2012 they were earning 49K. In other words, remove the impact of the real estate/LBO bubble of the late 80s and remove the impact of the credit bubble of the mid-2000s and solo practitioners incomes have, in fact, improved significantly over time relative to inflation.

But there are also problems with relying solely on the IRS data in GHF that go beyond this kind of cherry picking by those who choose to cite to GHF. While Professor Barton views the IRS data as the “best available” (in part, he explained to me, because it is longstanding and has been consistently collected over many years), it has its limits. This particular data set is, in fact, only a proxy for lawyers in solo practice and a proxy with some important qualifications.

1) Data Pollution. The data derives from an occupational category called “Legal Services” (described more fully here). It is not, in fact, a category that includes only lawyers. “Legal Services” includes numerous other non-lawyer occupations such as paralegals, patent agents, notaries, process servers and title officers. Thus, when it is said that there are 350,000 solo practitioners based on the fact that the IRS reports that approximately 350,000 tax returns were filed in the category Legal Services, we know that, in fact, the number of lawyers is smaller than that number. Because those non-lawyers likely earn less than lawyers it is likely that their inclusion skews the data towards lower incomes. In other words there are fewer solo practitioners than 350,000 and they earn more money than the data in the Legal Services category suggests (that is, more than 49K in 2012).

The IRS has confirmed this problem with me in an email from Emily Gross of the Statistics of Income Division of the Service: Our data follows NAICS codes and as you note, Legal Services includes both law offices as well as “other legal services”. “Other legal services” includes categories such as Notary public services, paralegal services, process servicing services, and real estate settlement offices. There is no way to separate Offices of Lawyers from the other data in the NAICS code 541100. The file is sorted according to these codes.

As an indication (but only an indication) of the potential impact of this mixing of classifications, the Census Bureau tracks the number of establishments categorized as “Offices of Lawyers.” They report there are only 174,000 such entities in 2012. These would include partnerships and incorporated entities. That result is, in fact, much closer to the data that the Census reports in its American Community Survey I discuss below for “self employed unincorporated” lawyers. The Census Bureau also reports a relatively small number of “other legal services” and the number do not add up to the 342,000 returns the IRS reports for the overall category. Unfortunately the IRS’s Ms. Gross confirms to me that there is no way to resolve this discrepancy with the IRS data.

2) Classification Mismatch. The data for filers in Legal Services is also only for “sole proprietorships” – a business entity that is neither incorporated nor a limited liability company. Sole proprietorships are not necessarily the same as solo practioners. It is possible since the 90s, and fairly common now, for solo practitioners to use a limited liability entity such as an LLC, a corporation or other form, in which to carry out their practice.

Census data I discuss below indicate a significant increase in “incorporated” self-employed lawyers in parallel with a decrease in unincorporated self-employed lawyers over the last 15 years.

If a solo practioner uses a limited liability entity, their earnings will not automatically appear in the IRS data presented by Barton.

A related problem is that while sole proprietorships are by definition owned by one person they can, and often do, hire other employees. A lawyer could hire associate lawyers, for example, as employees of the sole proprietorship. That means that some returns can show up in the Barton IRS data that do not represent true solo practitioners. The IRS data indicate that Lawyer sole proprietorships pay billions in salaries (not including salaries to the owner of the firm) suggesting the impact of these types of firms is not insignificant.

Further complicating matters, single member LLC’s (not the same as a solo practice as the LLC can hire associates who are not members) can elect to file a return as a “sole proprietorship” and more than a million LLCs do so every year across all occupational classifications. So those single member LLC’s could show up in the GHF data (and likely do) meaning the data overstate the number of solo’s.

Fortunately, the alternative Census Bureau data set I describe below provides a fix of sorts for some of these problems.

3) Gross Revenue v. Net Income. The “income” figures used in GHF and relied on by law school critics are what the IRS calls “net income” – in other words, the number is generated by the taxpayer him or herself after applying deductions to business receipts. (These returns are unaudited except in very rare circumstances.) The income for business receipts is, of course, much higher than the net income number.

For the tax year 2013, for example, “Legal Services” sole proprietorships generated approximately $40 billion in reported “Business Receipts” but only about $17 billion in net income. And in recent years that “Business Receipts” per return number has held up well against inflation. As an example, it was $120K per return in 2001 (measured in 2013 dollars) and $116K per return in 2013. That number is a much closer approximation to what the Census data I discuss below shows (the remaining delta between 116 and the higher Census results may be explained by the other issues described here). That seems to make sense – if the Census (like anyone else) asks you what you make each year you are likely to provide a pre-tax number that is akin to the “Business Receipts” gross revenue category on Schedule C for a sole proprietorship.

4) Cheating. Finally, of course, one has to ask, understandably, whether or not anyone should rely on the numbers that Americans provide the IRS. There are certainly well known incentives to under report income and over use deductions, in other words, to cheat. Given the very low rate of actual criminal prosecutions for tax fraud and the near certainty that one can reduce one’s tax obligations the incentives weigh on the side of under reporting income and over use of deductions. It turns out that lawyers as a class may be less likely to engage in outright cheating.

But it’s not really a question of whether or not lawyers cheat on their tax returns. It’s a question of whether they are capable of taking better (honest) advantage of deductions or other IRS rules that then help bias both the gross receipts and the net income numbers downward. That makes the IRS data a less reliable metric for anyone trying to understand the plight of solo practitioners. As one possible indicator, of the approximate 350,000 returns filed for sole proprietorships in the Legal Services category in tax year 2013 some 70,000 show no net income at all even though they generate more than 1.6 billion dollars in business receipts – somehow they found a way for all that revenue to disappear. (Data for TY 2013 and other years here.)

So, if the IRS data has problems, is there better data available? Well, the Census Bureau generates data that I think may provide a more accurate picture of what solo practitioners earn. It comes from the American Community Survey (“ACS”) (which goes back to 2001) and from the Decennial Census (“DC”) (for earlier years though limited to 10 year reporting periods). Mike Simkovic helpfully has extracted the relevant data from the ACS for all lawyers working full time (the vast majority) here and with both ACS and DC here (this time including part time lawyers and also breakdowns according to race and sex). Mike discusses some more detailed aspects of the ACS here.

This data has several advantages over the IRS data.

First, instead of relying on a category as broad as “Legal Services” the ACS data divides into two (relevant) classifications: lawyers who are self employed but unincorporated and lawyers who are self employed but incorporated.

This data, in other words, is lawyers only – no pollution from other occupations.

Further, it is a survey of individuals. When the two classifications are added up one gets about 300,000 individuals, or about 23% of the 1.3 million licensed lawyers in the US. That is higher than some estimates of the number of solo practitioners but lower than others. My guess is that many of those in the second category – self employed but incorporated – are not true solo practitioners because they use those entities to hire other lawyers who work for them – they are self employed meaning their own firm hires them but they, in turn, hire others as well. That may also explain, as we will see shortly, why the incomes in the second category are higher than in the first.

(The second category has one other complication – it only includes incorporated entities so it may not include LLCs. In fact, that may explain the bias downward in the Barton IRS data – practicing as a solo in a sole proprietorships is no longer the most sensible way to practice law and now that a variety of limited liability entities are available solo’s have shifted into that category which goes untracked in Barton’s data set. As I noted above it is tracked by the Census Bureau: in 1990 (just as the LLC form was being recognized along with other limited liability forms like the professional corporation) 213,000 unincorporated self-employed lawyers and only 77,000 incorporated but in 2014 there were only 167,000 unincorporated and 160,000 incorporated.)

As the ACS chart of all lawyers working full time (again, the vast majority) indicates, those in the first category (self employed and unincorporated – probably our best proxy for solo practitioners) have seen (mean/nominal) incomes grow from $116,000 in 2001 to $165,000 in 2014 (full time). And those in the second category (self employed but incorporated also full time) have grown from $151,000 to $186,000.

In buying power terms incomes in the first category rose from $155,000 to $165,000 and in the second fell from $201,000 to $184,000. Thus, it appears the second category has fallen behind. But a closer look indicates this is a bit misleading. That category has more volatile results and if one uses 2013 as a benchmark (when their income was reported as $197,000) those lawyers keep even with inflation.

Of course, the results in both categories fall when part time lawyers are added. The impact, though, is relatively minor. Instead of staying ahead of inflation over the last fifteen years, for example, those in the unincorporated category stay dead even. You can also go back to 1990 (with the help of the Decennial Census data) and you find that lawyers in that same category stay even with inflation.

Keeping even with inflation might not seem like an achievement but it means that relative to how others are doing in an era of an allegedly stagnant middle class (a controversial proposition), lawyers in solo practice are holding their own. Much of the critics’ argument that law school is in need of a drastic overhaul has been based on the idea that law schools are producing graduates unable to hold their own even relative to less well educated workers. There seems little basis to this conclusion given this data as opposed to the anecdotes that many of the critics rely on.

And, of course, holding your own as a lawyer at an annual income of $147,000 or $165,000 or $184,000 puts you way ahead of most of the country’s workforce (where the mean household income is about $73,000).

That means there is little basis to the case that some kind of tectonic shift in the lives of lawyers has taken place. It suggests that for lawyers, at least, the glass is more than half full. It also implies that whatever their ups and downs in the market place law schools are continuing to produce trained professionals who are in demand in the marketplace.

13 Comments


  1. One quick note about that California bar survey which I have now (briefly) reviewed: it only provides three categories: partners, associate, and “solo proprietor” (sic). So, what are we to make of that? Embarrassing certainly. But presumably any equity owner in a PC would have called themselves a “partner.” And I am not sure what a genuine solo practice lawyer would have done as there is, of course, no such animal as a “solo proprietor.” So I think we have to view this survey with caution.


  2. Thank you for your continuing thoughtful comments, Gillian.

    I agree we need more precise data but I think that depends on what it is we want or ought to be asking.

    This debate began because Ben published data suggesting what appeared to be a rather startling collapse in a class of lawyers that he maintains represents the core (“middle class”) of the profession.

    Beyond the somewhat imprecise manner of the presentation of the data, it turns out that if that collapse occurred, it likely occurred a long time ago and is likely not terribly relevant to what is going on in the profession today. It is therefore also far less likely to be relevant to what is going on inside law schools today. As the work of the American Bar Foundation demonstrated with their original Chicago survey, the really important shift in the profession impacting solo practices occurred in the 70s and 80s when the large firm first emerged in a serious way. True solos who owned sole proprietorships declined dramatically then as new lawyers increasingly joined law firms and as solos formed newly available limited liability entities.

    Lawyers who form those entities can, of course, still be solo practitioners but the large number of those types of entities means if you really feel it is important to examine solo practices you have to find out what is happening to those kinds of lawyers. You can’t just focus on sole proprietorships which is what Ben does as long as he relies on IRS data alone.

    The former development (shift away from true solo practice/sole proprietorships) though also means that it may not be as important as it once was to examine what is happening to true solo practitioners because it is simply not the most important way for lawyers to earn a living.

    Of course, it is still important and arguably in certain areas it is more important than it may have been a few years ago. One development I see here in the Valley (where there is a kind of natural experiment available to anyone interested because of the emergence of a new generation of large firms in the tech sector) is the spin-off of experienced lawyers from the large firms who generate significant work from the conflicts faced by those large firms. In areas like IP and employment law it is feasible to build a successful solo or small firm practice (although far less feasible to do so in areas like public company work or large financial transactions.)

    In other words, I think the real problem with Ben’s approach is that he latched on to a widely held but dated Atticus Finch image of lawyering and then, when he easily established that it was not doing so well, came to a somewhat unsustainable conclusion.

    I don’t contest necessarily that there are 300,000 or so “solo practitioners” – only that I doubt that more than a certain limited percentage are in sole proprietorships. I think it is instructive (as you discuss) that if you add the two ACS categories together you get pretty close to the number of solos you and Ben think exist. Sure, there are some partners in the mix but how many? I have not seen the Cal Bar data but are all of those folks genuinely “partners” as defined by California law?

    Most firms beyond a minimal size surely organize as a PC, PA, LLC or LLP. Only the latter is likely to lead to partner income being reported in the first ACS column. And as you note the sole proprietorships can hire employees, including associate lawyers. That is another problem with the reliance on the IRS data.

    In any case, I also return to a comment I made much earlier in this discussion: the ACS reported incomes are much closer to the “business receipts” numbers that Ben’s IRS data on sole proprietorships indicates (that is, prior to deductions). I think that this is likely the best explanation for the gap – it just doesn’t make sense to me for any lawyer to say, when asked, oh, what I earn is what I report as my net income to the IRS after I take out deductions for my home office. That’s like me saying I earn $X pre-tax from Santa Clara after deducting my health care contribution and my pension contribution . I think I would likely gain a lot more sympathy from the scam critics if they viewed my academic salary in the same way they want to view the earnings of solo practitioners.

    So, when a solo is asked by the ACS what she makes she says, $120,000 but her tax return shows “net income” of $50,000. That doesn’t require cheating it’s just two answers to two different questions.


  3. I think we can all agree on the following: none of these data sets identifies all solo practitioners operating law offices and estimates their income reliably. And we should probably also all agree that the estimates we care about depend on what one cares about measuring–the income of a solo practitioner who works full-time in this and only this job, the income of any one who practices solo, full or part-time (or maybe only those part-timers who are involuntarily practicing less than full-time or those without additional income), or any small practitioner–solo or small partnership–who is serving primarily individual and small business clients. So we should all be on the same page that understanding more about the economics of solo/small firm practice is a matter of judgment and triangulation from multiple, imperfect, data sources.

    Stephen (please call me Gillian) raises a couple of issues about which I have some ideas. First, I agree with Ben that a good estimate of the number of solo practitioners is about 300,000 (or a little more.) The ABA in 2012 reports about 1.2 million licenses in the U.S. The ACS data count 1 million “lawyers, judges and other judicial workers.” There are about 30,000 judges, state and federal, plus some number I don’t know of administrative law judges, magistrates, clerks, etc. So let’s call that total 100,000, giving us about 900,000 lawyers. The ABF’s Lawyer Demographics for 2005 (last year available) estimates that about 75% of all license holders are private practitioners and about 50% of private practitioners are solos; the 2011 California State Bar Survey (1,800 sample, all license holders active and inactive) found 67% of license holders who answered yes to the question “do you currently work as an attorney” were in private practice and of those in private practice and currently working as an attorney, 47% were solo practitioners–that was 33% of all license holders who said they were working as attorneys. (Note this number has bounced around–in the 2006 survey, 40% of attorneys worked solo, in 2001, 35%.) So, given those multiple surveys, a rough estimate of one-third solo practitioners among all working as lawyers is not a bad guess. That makes an estimate of the number of solo practitioners of 300,000 also not a bad guess. We still don’t know how many of the IRS sole proprietorships in legal services are solo practitioners–not all of them for sure; and that category doesn’t count solos who form a professional corporation/LLC–and those entities are not available in all states and tax advice on this varies.

    So that leaves the question Stephen asks of where are all these solos in the Census data? In the ACS, the answer is just not clear, but the question I have is, where are all the partners? The “self-employed” incorporated and unincorporated category, he says (I haven’t seen the data myself, just the published tables for the five-year sample 2006-2011 available at ACS) add up to about 327,000. If those categories include partners (which I’m very sure they do, of all income levels, but not consistently with the intention of the survey since there’s probably substantial reporting and estimate error here–all lawyers other than those in PCs should be here for tax purposes) then the number is too low. Not having seen this dataset, I’m not sure what other categories may contain partners and what those numbers are. If we used the Cal Bar study as a basis for estimating, there should be about 180,000 partners (Cal found 30% of those in private practice are partners so if 2/3 of the 900,000 lawyers estimated by ACS are private practice that gives 180,000).

    As for the higher average for self-employed incorporated as opposed to self-employed unincorporated, two things to say: first of all, I don’t think there is any a priori way to say which is more likely to contain solos as opposed to partners (and remember, the vast majority of partners are in small firms–75% of all law firms are 2-5 partners; most but not all of these will be serving the individual market and not earning mega-incomes.) Tax advice goes both ways. But I think one could predict that those with higher income are more likely to incorporate as they have more to protect against liability. Overall, it’s really important to remember that there is massive skew to the distribution of lawyer earnings, so means can be meaningless. And while the vast majority of solos make low income, this is also the category in which you can find some of the most highly remunerated lawyers–PI lawyers, for example.

    Another observation: the counts in the Economic Census of establishments suggests a large number of solos must be very small–with no payroll. The Economic Census only counts law firms with payroll (employees). There were 171,000 of those in 2012. 28% were individual proprietorships, 17% were partnerships, 35% S corporations (owner/partners should be classifying themselves for IRS purposes and probably ACS as self-employed and taking most income as salary, but they’re not) and 15% were C corporations. Any way you slice it, there must be lots and lots of solos who are not in the data–that is, don’t have payroll. Some of these will be working a few hours as a sideline to other employed work; others will be working low hours, and that’s their only source of income.

    Last comment: the ACS data for 2006-2011 estimate that 46% of “lawyers, judges, judicial workers” made less than $100,000, 32% less than $75,000. Average judicial salaries are $150,000 or more; magistrates probably make an average of about $100,000; judicial clerks (permanent) maybe around 40,000. So it’s hard to know how they impact the averages but conceivably not at all. In the Cal survey, 26% of lawyers reported earning less than $50,000; 48% less than $100,000 from their legal practice (includes employed lawyers who say they are working as attorneys.) Of those who said they worked as attorneys, 21% said they worked less than 35 hours (that is, less than full-time) in their legal practice. So even being conservative–assume all the low hours people are also low earners (needn’t be true–a former big law partner now working solo for a few clients could be part-time and still well-paid), that indicates about 30% of lawyers working full-time as attorneys earning less than $100,000. And a lot of those folks are solos.


  4. Professor Barton has weighed in again and now agrees that indeed the data he relied on for his book to demonstrate the weakening earnings of solo practitioners was inaccurately described. Presumably in future editions of the book and other publications he and others who have relied on this data will note this problem.

    Professor Barton now suggests (as does Professor Hadfield in her comments here) that there is a possible problem with the ACS data I refer to as a proxy for solo practitioners because this data may also include partners. The ACS data includes two categories of “self employed” lawyers, one that is for lawyers who incorporate and one that is for lawyers who do not incorporate. The latter category is the one in which it is conceivable that “partners” of law firms are included. This, it is argued, distorts the incomes reported in that category upwards and thus we lose a clear picture of what is happening with solo practitioners.

    Maybe but maybe not.

    If, in fact, that ACS category is where the partners at Cravath or Skadden report their incomes then one would presume there would be an upward pull on the reported incomes. But Cravath – like almost all major law firms – is a limited liability entity. Some limited liability entities can elect to be treated as corporations for tax purposes and thus would presumably fall out of the Unincorporated ACS category and into the Incorporated ACS category. (Professor Hadfield notes a U.S. Census Bureau publication suggesting that the ACS is driven by IRS categories.)

    This is actually consistent with the glossary of terms that Professor Barton refers readers to in his latest post. This glossary is used by the Census Bureau when it conducts a special survey for the city of New York and is said by the Bureau to be “close” to the definitions used by the ACS. Here is the wording:

    “Self-employed in own incorporated/unincorporated business or professional practice.

    Own business, incorporated, refers to people who own all or most of the stock in a privately held corporation, and consider themselves self-employed.

    Own businesses, unincorporated, refers to work for profit or fees in the person’s own business, shop, office, etc. It does not include managers or other executives hired to run a business, salespersons on commission, or corporate officers. This category includes sole proprietorships and partnerships, but the company cannot be incorporated.”

    It appears that the first category would include Professional Corporations (many law firms are PCs – like my old firm Wilson Sonsini Goodrich & Rosati) but may or may not include LLCs (there is no “stock” per se in an LLC since it is a “company” not an incorporated entity but there can be units or membership units and of course as noted an LLC can elect to be taxed as a “corporation.)

    So it is not clear that, in fact, the lawyers at the large law firms where partner incomes would presumably drive up the ACS earning data are in fact in the first category.

    There is a larger problem with this line of attack from Professors Hadfield and Barton. If the presence of “partners” was having a significant impact on the data, then it is hard to understand why the incomes in the “self employed unincorporated” category are LOWER than those in the “self employed incorporated” category. Either the presence of “partners” is not that significant (not because they do not earn more money but because there are not enough of them) OR they are in the second “incorporated” category where incomes are indeed higher because they are only metaphorically partners and are really part owners of limited liability entities like PCs or LLCs (and perhaps even LLPs if they are viewed as corporate like “entities” not pure pass through vehicles). I noted some of these issues in my original post.

    A further problem is that Professor Barton can’t seem to find all the solo practitioners he needs to find. He says there are 350,000 of them, more or less. But the ACS only gets close to that many (327,000) when you add up both self employed classifications (incorporated and unincorporated). It might be reasonable to conclude that, in essence, there are more than 300,000 solo practitioners (although that seems too high) if we just agree to disregard the choice of entity that solo’s use. But if we want to do that then we have to accept that their incomes – at least according to the ACS – are far higher than the original “sole proprietor” data presented by Professor Barton indicates. That data included non-lawyers and did not track incorporated entities (which the ACS says account for half the self employed lawyers).

    A far more likely explanation for the disconnect is that Professor Barton relied on a form of legal practice – solo practitioners operating sole proprietorships – that was dying out in the face of the emergence of new alternative limited liability forms for operating a legal practice such as LLCs, PC’s and LLP’s. I explored this in my original post since the increase in the “incorporated” category correlates with the wider recognition and acceptance of these new entities and the flexibility they provide to those who use them. The ABA even provides advice to solo’s about how to use these entities. It is clear from the ABA piece that they discourage use of the sole proprietorship form so it is not a surprise that the most enterprising and successful solo’s have abandoned it. But that means they are not present in the Barton IRS data either.

    This points to the final problem I will note with Professors Barton and Hadfield’s latest line of defense: it is possible to operate as a single member LLC or singe shareholder PC. As the ABA article linked above suggests those are among the best alternatives if one wants to gain limited liability protection. (Actual mileage may vary as not all states recognize all of these forms. California allows lawyers to form PCs and LLPs, for example, but not LLCs.)

    In fact, one could go back to Professor Barton’s original IRS data. As he notes the IRS collects income data for both sole proprietorships and partnerships. He presumes (with minor caveats that he largely dismisses) that “solo practitioners” are “sole proprietorships.” But that is not true. In fact, that is one reason for the attraction and explosive growth of the new limited liability entities. (There are now close to 9,000 “law corporations” in California alone and in a survey done in 2011 nearly 50,000 nationally (half the total of law firms that were not sole proprietorships.)) Solo practices can be set up using these and their incomes are likely not be showing up in his data.


  5. Correction to my post above [below SD]: the Economic Census uses administrative data for small firms, not the ACS. And: it does appear that some LLCs can file as S-corps–but I’m not claiming any tax expertise here. The key conclusion is: there clearly are partners in the self-employed category (and almost all probably should be) and there’s no reliable way that anyone has yet identified for distinguishing solo from partner using incorporation as the marker. Incorporated tend to be higher income, according to Mike and Stephen, but that could be reverse causality: those with more income see more tax benefits to incorporation.


  6. A follow up on the American Community Survey which Mike and Stephen are looking at. The “self-employed” category on that survey includes all lawyers in private practice–including partners in all size partnerships, small to huge–with the possible exception of those who are organized as professional corporations (which LLCs and LLPs are not.) See https://www.census.gov/prod/2013pubs/acsbr11-21.pdf for some discussion. Even Cravath LLP lawyers are self-employed for IRS purposes, which is what the ACS uses. In California, there are approximately 8,800 professional corporations (LLCs are not available as a form for lawyers) and about 190,000 active practitioners. Tax advice to practitioners in California advises partnerships to organize as LLPs not professional corporations; and advises solos not to incorporate unless they have concerns about liability for their employees (they can’t shield themselves from their own negligence liability by incorporating). http://www.staleylaw.com/images/Choice_of_Entity_for_Practicing_Law_in_California_-_48395.pdf. A tax colleague of mine suggests almost all lawyers in private practice are self-employed for tax purposes (and hence for ACS purposes) and the only ones who might see an advantage to forming an S-corp (which LLCs are not but professional corporations are) are possibly solos and those with super-high income. I think it’s very difficult to determine who’s in the incorporated versus not incorporated category (and no doubt there’s significant error in census responses; but note that for small firms, ACS may be relying not on survey answers but administrative records, that is, things like the IRS for income).

    But the bottom line is, I don’t think this data source can help us determine the earnings of solo practitioners.


  7. I think Ben’s numbers for current average income are probably closer to the truth, although the biggest message here is the data are crap and we should not be making the types of guesses we have to make to know this basic fact about a key aspect of legal infrastructure.

    I’ve looked at the Economic Census for 2012 for Offices of Lawyers. The bottom 40% of law firms (57,612 establishments) reported annual receipts averaging $134,403 and average payroll of $47,633. That leaves $87,773 after payroll–out of which all other expenses (office rental and equipment, insurance, etc.) are paid. This implies a take-home before tax that could easily be about $50,000. And certainly is (probably much) less than $90,000. The data are apparently drawn from other federal agencies for very small firms so this would include payroll data reported to the IRS as well as sales data reported for purposes of business licenses as well as federal and state taxes. The 40% cutoff (which is all law firms earning less than $250,000) probably overcounts solo practitioners but not by much (and so overstates solo pay) and it doesn’t vary with business type (incorporated or not). It doesn’t include other legal services.

    And: it looks like things might have gotten worse after the recession. The 2007 Economic Census gives shows (2012 dollars) $146,981 in receipts for bottom 40%, $48,831 in payroll for net receipts of $98,150. That’s for 60,690 establishments. Because the cutoff I’ve used in both is that available in this series–firms earning less than $250,000 (in current year dollars) there’s no easy way to determine how many small firms went out of business and so the extent to which the average of the surviving firms fails to capture the effect of the recession.




  8. I would add to the list of questions for Ben the following:

    1) I think we all agree now that the original IRS data in your book was (clearly unintentionally) mislabeled “solo practitioners” when, in fact, the data base was for a larger category called Legal Services that includes, as the IRS explained to both of us, non lawyers. Since the IRS has also explained that their data does not allow us to break down the category into lawyers and non-lawyers we do not know what proportion of the 342,000 returns (in the most recent tax year) are for lawyers. I did not claim and do not believe that half of the figure are for non-lawyers, but I do know some are for non-lawyers and those non-lawyers earn less money than lawyers and that that generates a lower net income figure. Therefore, should we not conclude that we cannot rely on the IRS data alone to tell us what is happening to the incomes of solo practitioners?

    2) While you continue to rely on the IRS data to back up a conclusion that solo practitioners are doing much worse than they have before there is also another possible explanation, isn’t there? Since the early 90s the number of self employed but incorporated lawyers has increased dramatically as tracked by the ACS data. Those incorporated lawyers have significantly improved their earnings over time as well. Those incorporated lawyers can be solo practitioners and many likely are. Meanwhile, by definition, the solo practitioners in the IRS sole proprietorship category are not incorporated. It is conceivable that those are the lawyers who got “left behind” – that is, the higher earning more successful solo practitioners were willing to spend the time and money to incorporate. Their higher earnings show up in the ACS data but not in your IRS sole proprietorship data.

    In any case, clearly this remains contested space. I certainly welcome any additional contributions you might have to the discussion.


  9. Mike Simkovic poses several as yet unanswered questions to Ben Barton about his reliance on IRS data here.


  10. Mike Simkovic has weighed in on this discussion here.


  11. Thanks for your interest in my work! I have several responses:

    1) I disagree with your analysis of how many of the returns covered in the IRS data come from lawyers and how many come from “other legal services.” As you noticed in your back and forth with the IRS data people, it is not possible to tell exactly how many of each group is counted in that NAICS code and there is not a way to square the IRS and Census data based on what we know. I think the IRS data includes more lawyers than non-lawyers, because there are more than 4 million notaries in the US (let alone the various other people that are included in “other legal services”), so if every non-lawyer sole proprietorship was included in the IRS data the number of filers included would be much higher and mostly notaries/other legal services. Also, the most recent ABA count of solo practitioners is relatively close to the IRS count (combine http://www.americanbar.org/content/dam/aba/administrative/market_research/lawyer-demographics-tables-2016.authcheckdam.pdf and http://www.americanbar.org/content/dam/aba/administrative/market_research/total-national-lawyer-population-1878-2016.authcheckdam.pdf) so the ABA numbers are close to the IRS numbers. Likewise, if you just add the two census categories of lawyers you discuss in your post that number is also close to the IRS number. Because the IRS data cannot be squared with the Census data I think my guess is as good as yours here.

    2) Based on your numbers, I think that you think that only about half of the IRS tax returns come from lawyers, but maybe that is wrong. Assuming that is right, you use a census count that comes to around 174,000 in 2012. The total number of filers in the IRS data set is around 350,000 in that timeframe, so for the rough calculations below assume that only half of the IRS data set comes from lawyers. You also think that these lawyers earn an average (mean) of $165,000 a year. That is hard to square with the data. Why? The average earnings in the IRS data was below $50,000 during every year from 2010-2013, but let’s just use $50,000 for simplicity’s sake. If half of a group of people earn $165,000 a year (and the incomes within that average are not distributed weirdly), the other half of those people must earn -$65,000 a year to have an average of $50,000 across the entire group.

    (How did I calculate that? 174,000 is roughly half of the sample size. If half the sample size averaged $165,000 a year, the other half must earn an amount that makes the total average roughly $50,000. In order to do that the sum of $165,000 + X must equal $100,000, since 100k/2= 50k. Admittedly the 165k may be distributed weirdly, with a thin layer of very large earners and a large group of lower earners. If that is the case I do not think that helps you).

    3) Regardless, we agree (I think) that the IRS has used the same process and the same data collection since 1967. I see that you do not like the year 1988 as the comparison point. That is fine. Chose the first year (1967), or really any of the years between 1967 and 1972, and you will see that filers in this category have lost significant ground over the last generation. Since the beginning of the data these folks have lost roughly a third of their real purchasing power. I do not think that loss of earnings comes solely from whatever amount of non-lawyers are in the sample.

    4) I think you suggest that lawyers may report their gross revenue to the Census and their earned income (after deductions) to the IRS. I think that is likely as well, which makes the Census data less reliable as a measure of how lawyers are doing. Income is the relevant question, not gross receipts.

    5) The IRS data is subject to criminal penalty if falsified, including both fines and potential jail time. With the possible exception of accountants, no other group of Americans is more aware of this fact than lawyers. The Census data is collected anonymously, often off the top of the person’s head while filling out a government form.

    6) The IRS sole practitioner data on shrinking incomes matches the most recent After the JD data, other bar surveys, and the NALP data that shows a tough employment market at the low end for lawyers since the 1990s. By some calculations roughly a third of all JD holders since 1990 were not able to secure work 9 months after graduation that required a JD. I argue in my book that the NALP data matches the IRS Data – something changed for small firms and solo practitioners around 1990, and main street lawyers steadily lost ground from then until now.

    7) I know that you and others feel very, very strongly that everything in the previous paragraph is wrong and that you are “all in” on the census and BLS data. That is fine and justifiable, although, like your reaction to my data, I have some reservations about your data, much of which has been covered by others elsewhere. Just as you have stuck to your guns with your data, I prefer the IRS data, so hopefully we can pleasantly agree to disagree.

    8) Thanks for your kind words on my data collection efforts, for discussing my book and for linking to it in the post.

Comments are closed.