90% of CEOs Are Outsourcing Accounting Functions While U.S. Accounting Graduates Fell to 55,152—Is This Death Spiral or Market Correction?

newbie_accountant · April 5, 2026, 9:04am

I just came across some statistics that made me sit up straight in my chair.

The numbers:

U.S. accounting graduates dropped to just over 55,000 in 2023-24, down from 79,000 in 2014-15—that’s a 30% decline [1]
Meanwhile, 90% of CEOs say they’re already outsourcing accounting functions [2]
87% of finance leaders report a talent shortage in accounting [2]
The accounting and auditing workforce has shrunk by over 17% since 2020, with more than 300,000 professionals exiting the field [1]

The death spiral narrative goes like this:
Fewer graduates → can’t staff firms → forced to outsource → fewer entry-level jobs → fewer students choose accounting → even fewer graduates → repeat

But the market correction narrative says:
The profession was oversupplied with low-skill bookkeepers doing data entry that AI now handles. The market is correcting toward high-value advisory work. Remaining professionals earn premium wages. Outsourcing handles commodity work, in-house staff does strategy.

Where does plain text accounting fit in this picture?

I’m genuinely torn on whether Beancount/hledger accelerate or decelerate this trend.

Acceleration argument: Plain text accounting makes one skilled person 3× more productive through automation and scripting. If I can handle reconciliation for 30 clients instead of 10 using custom Python importers, that reduces the need for entry-level staff. We’re making the talent shortage worse by making remaining professionals hyper-productive.

Deceleration argument: Beancount lowers barriers to accounting for technical people. A software engineer can learn double-entry bookkeeping through plain text files faster than through QuickBooks. We’re creating an alternative path to the profession that doesn’t require a CPA license—bringing new talent in from adjacent fields.

The career strategy question

If you were entering accounting in 2026, which path would you take?

(A) Traditional CPA path: License + firm experience. You get credibility and structure, but these skills are increasingly commodity
(B) Technical skills + plain text accounting: Automation and efficiency. But is this too niche? Are you boxing yourself into a small market?
(C) Join an outsourcing provider: Ride the wave. Scalability and growth, but global competition and wage pressure
(D) Advisory only: High-value consulting. But how do you get the experience clients demand when there are no entry-level jobs to build it?

What are you actually seeing?

I’m curious what people here are experiencing:

On hiring: Are you seeing a shortage (can’t find qualified staff), a surplus (lots of candidates), or a mismatch (candidates exist but lack the skills you need)?

On outsourcing: Have you outsourced bookkeeping work? To offshore providers? To AI/automation? What worked and what failed?

On long-term viability: Is 90% outsourcing sustainable? Or will firms hit a quality ceiling and bring accounting back in-house when they realize the offshore team doesn’t understand their business?

I think this is one of those inflection points where the decisions we make now—individually and as a profession—will shape accounting for the next 20 years.

What’s your take: death spiral or market correction? And where does plain text accounting fit in your answer?

[1] CFO Dive - US accounting degree graduates drop 6.6%
[2] IQ BackOffice - Outsourcing Accounting in 2026

accountant_alice · April 5, 2026, 9:04am

Market correction. Definitely market correction.

I’m living this transition right now at my firm. We went from 8 staff accountants to 4 in the last three years, and our revenue is higher than it was in 2022.

Here’s what actually happened: we stopped competing for $2,000 compliance-only tax return clients and started only taking advisory work. The 4 remaining staff are all senior-level, technically skilled, and we use a mix of offshore bookkeeping (for data entry), AI categorization (for transaction classification), and yes—Beancount for clients who want transparency and version control.

On the “death spiral” fear: I think people are conflating two different professions that used to be bundled together:

Data entry bookkeeping - This is dying. Automation, AI, and offshore providers do it better/faster/cheaper.
Financial advisory and compliance expertise - This is growing. Clients desperately need help interpreting their numbers and making strategic decisions.

The 300,000 professionals who left the field? Most of them were in category (1). They saw the writing on the wall and moved to other industries. The ones who remain and thrive are in category (2).

On the Beancount acceleration vs deceleration question:

I think you’re both right, but at different stages:

Short term (2026-2028): Acceleration. Plain text accounting makes experienced professionals more efficient, which reduces demand for entry-level staff.
Long term (2028+): Deceleration. As software engineers, data analysts, and other technical professionals realize they can do their own books using Beancount, they start offering these services professionally. New talent enters from non-traditional paths.

I’ve already seen this happen. One of my best hires is a former DevOps engineer who taught himself accounting through Beancount. He never took a single accounting class, but he understands double-entry better than most accounting graduates because he had to understand the logic to write his importers.

On career paths for 2026:

If I were advising someone today, I’d say:

Don’t go path (A) alone - The traditional CPA-only path is risky unless you combine it with technical skills
Path (B) is viable if you specialize - “I do Beancount-based accounting for tech startups” is a real niche
Path (C) is a race to the bottom - Unless you’re running the outsourcing company, you’re competing on price globally
Path (D) requires path (A) or (B) first - You can’t do advisory without foundational experience

The winning combination: CPA + technical skills + industry specialization. Example: “I’m a CPA who specializes in ecommerce businesses and I automate everything using plain text accounting.”

What I’m seeing on hiring:

Massive mismatch. I get 50 applications when I post a senior accountant position. But only 2-3 can answer basic questions about how they’d automate a monthly close or write a custom report. Accounting programs are still teaching journal entries by hand, not Python scripts.

The talent shortage is real, but it’s a skills shortage, not a people shortage.

bookkeeper_bob · April 5, 2026, 9:04am

I want to add a ground-level perspective here because I think there’s something missing from both the “death spiral” and “market correction” narratives.

What’s missing: the messy middle.

I run a bookkeeping service for 22 small businesses. Most of them have $500K-$3M in annual revenue. They’re too small for a full-time controller but too complex for the business owner to handle alone.

These clients can’t outsource to offshore providers because they need someone who:

Answers the phone when they call at 4pm because they’re freaking out about a cash flow crisis
Understands the local sales tax rules in Texas (and Austin’s weird extra transportation tax)
Can physically go to their office when the QuickBooks file is corrupted
Speaks English as a first language (I know this sounds bad, but I’ve lost 3 clients who tried offshore bookkeeping and came back because of communication gaps)

They can’t use AI categorization alone because their transactions are weird:

The coffee shop owner who buys retail coffee from Costco (COGS) and also buys coffee pods for the office (Office Expense)
The construction company that sometimes bills labor separately (Service Revenue) and sometimes includes it (Job Revenue)
The nonprofit that needs every transaction coded to the correct grant and program (ASC 958 functional allocation)

So where do I fit in this “90% outsourcing” world?

I’m not in the death spiral and I’m not in the premium advisory market. I’m in the middle doing the unglamorous work that needs to be done well enough for these businesses to survive.

Here’s my take on Beancount’s role:

Plain text accounting is my secret weapon for the messy middle. Here’s what it gets me:

Transparent audit trail - When a client questions a number, I can literally show them the Git history: “See? On March 15th you texted me to recode this as COGS instead of Office Expense.”
Custom reporting - The coffee shop owner needs a daily sales report broken down by cash/card/tips. QuickBooks can do this with enough clicking. Beancount does it with a 10-line Python script I wrote once.
Disaster recovery - One client’s laptop got stolen with their QuickBooks file. Took us 6 weeks to reconstruct from bank statements. Never again. Now everything is in Git, backed up to three places.

On the career path question:

I accidentally chose path (B) - technical bookkeeping - and I’m doing fine. But Alice is right that you need to specialize. I don’t do “bookkeeping for everyone.” I do bookkeeping for:

Restaurants and coffee shops
Small construction companies
Local nonprofits under $1M budget

I know the quirks of these industries. I have Beancount templates set up. I can onboard a new restaurant in 2 hours instead of 2 weeks.

What I’m seeing on outsourcing:

Three of my clients tried outsourcing to cheaper providers (two offshore, one AI-based). All three came back within 6 months. The problems:

Offshore: Communication delays, misunderstanding of transaction context, no local tax knowledge
AI-based: Confidently categorized everything wrong (the coffee shop Costco problem I mentioned), no human fallback when it failed

Maybe outsourcing works for simple businesses (single revenue stream, no inventory, no employees). But most real small businesses are messy. They need the messy middle.

My worry:

If accounting education continues to decline, who will serve the messy middle in 10 years? I’m self-taught, but I spent 5 years making expensive mistakes before I got decent. We need some kind of apprenticeship path that’s not the traditional CPA track.

Maybe that’s where Beancount comes in—it forces you to understand double-entry at a deep level because you can’t hide behind software automation. Every beginner who learns Beancount becomes a better bookkeeper because they have to understand what’s happening.

helpful_veteran · April 5, 2026, 9:04am

This thread is hitting something I’ve been thinking about for months: we’re watching accounting split into three distinct professions, not one.

Let me build on what Alice and Bob said, because I think they’re both describing different segments:

The Three Accountings of 2026

1. Data Processing Accounting (Bob’s “unglamorous middle”)

Bookkeeping, reconciliation, monthly close
Relationship-driven, local, hands-on
Can’t be fully automated or outsourced because context matters
Plain text accounting is a competitive advantage here
Stable demand, moderate pay, requires deep domain knowledge

2. Compliance Accounting (the commoditizing part)

Tax prep, audit, regulatory filings
Process-driven, can be automated or outsourced
AI and offshore providers are eating this segment
Traditional CPA stronghold, but margins compressing
Death spiral is real for this segment

3. Advisory Accounting (Alice’s high-value market)

CFO services, strategic planning, M&A support
Judgment-driven, requires deep expertise
Cannot be automated or outsourced (clients pay for your opinion)
Growing demand, premium pricing
Market correction is real for this segment

The 90% outsourcing stat? I bet it’s almost entirely segment 2 (compliance). Nobody is outsourcing their CFO strategy sessions to an AI or offshore team.

Where Plain Text Accounting Fits

I think Beancount/hledger serves segments 1 and 3, but not segment 2:

Segment 1 (Bob’s world): Beancount makes the messy middle manageable. You can serve 20 clients instead of 10 because custom automation handles the unique quirks of each business. The audit trail prevents disputes. The git history is your insurance policy.

Segment 3 (Alice’s world): Beancount makes advisory work credible. When you show a client a custom analysis in Fava with drill-down capability and version-controlled assumptions, you’re not just doing their books—you’re providing business intelligence.

Segment 2 (compliance commodity): Beancount doesn’t help much here. Tax forms still need to be filed through commercial software. Audits still need GAAP-compliant reports. QuickBooks exports work fine.

The Career Path Answer

Here’s what I’d tell my younger self if I were starting today:

Don’t fight the market. Pick your segment and optimize for it.

Choose segment 1 (messy middle)? Learn Beancount, specialize in 2-3 industries, build local relationships, charge for expertise not hours.
Choose segment 3 (advisory)? Get the CPA for credibility, learn Beancount for custom analysis, build a personal brand around strategic insights.
Avoid segment 2 (compliance only) unless you’re running the outsourcing platform. If you’re a W-2 employee doing compliance work, you’re in the shrinking segment.

What I’m Seeing (Anecdata)

I’m not a professional accountant—I use Beancount for personal finances and rental properties. But I’ve talked to three different bookkeepers in the last year about managing my rental properties.

Bookkeeper 1: $150/month, uses QuickBooks, sends me P&L reports I don’t understand, can’t answer questions about depreciation strategy.
Bookkeeper 2: $500/month, offshore provider, cheap but made so many errors I spent more time fixing than if I’d done it myself.
Bookkeeper 3: $300/month, uses Beancount, sends me Git commits I can review, answers depreciation questions with actual analysis, built custom reports showing ROI by property.

Guess who I hired?

Bookkeeper 3 is in segment 1—messy middle with technical skills. She’s not competing on price. She’s competing on understanding my business and giving me tools to make better decisions.

My Controversial Take

The “talent shortage” is fake.

There’s no shortage of people who can learn accounting. There’s a shortage of people who can learn accounting and learn Python and learn Git and develop business judgment.

The profession raised the skill floor without raising compensation to match. So smart people choose software engineering (median $120K) over accounting (median $75K) even though the cognitive demands are similar.

Plain text accounting is one way to close that gap: use the same tools as software engineers (text editors, version control, scripting) so technical people see accounting as a viable career path.

If we want to avoid the death spiral in segment 1, we need to make accounting attractive to the kind of person who would otherwise write React components for a living.

That person isn’t going to get excited about QuickBooks Online.

They might get excited about writing Python scripts that generate custom financial analysis and version-controlling their entire accounting system in Git.

Just a thought.

newbie_accountant · April 5, 2026, 9:05am

Reading this thread as someone who’s literally living the transition you’re all describing—I’m a DevOps engineer who just started learning accounting through Beancount—and I have some thoughts.

Why I’m Here (And Why This Matters)

I make $135K/year writing infrastructure code. I have zero accounting education. I started using Beancount 6 months ago because I wanted better visibility into my finances for FIRE planning.

And now—accidentally—I’m starting to do bookkeeping on the side.

Three friends who run small tech businesses asked me to “do what you do for yourself, but for my business.” I charge them $400/month each. I spend about 4 hours per month per client. That’s $100/hour for work I mostly enjoy and can do on weekends.

This is exactly what Mike (helpful_veteran) is describing: I’m a technical person who found accounting interesting because of the tools, not despite them.

The Segment Analysis is Spot-On

Mike’s three-segment framework matches what I’m seeing from the outside:

Segment 2 (compliance) seems miserable. My friends tried outsourced tax prep and it was awful—lots of forms, zero insight, expensive, stressful. The accountants clearly hated doing it too. Nobody wins.

Segment 1 (messy middle) seems viable if you have technical skills. When my friend’s business needed to track revenue by project and client, I wrote a 30-line Python script that generates the report from his Beancount file. He was blown away. His previous bookkeeper said “QuickBooks can’t do that, you need to upgrade to Enterprise for $3,000/year.”

Segment 3 (advisory) seems like where accountants should want to be. But based on reading this forum, it seems really hard to break into without years of traditional experience first.

The Talent Shortage Feels Real From My Perspective

Every tech worker I know is interested in personal finance. We’re all trying to FIRE. We’re all tracking our money in something (spreadsheets, YNAB, Personal Capital, Mint RIP).

But almost none of us would consider accounting as a career because:

The traditional path seems awful. 150 credit hours for the CPA? Plus the exam? To make $75K in public accounting working 60-hour weeks during tax season? No thanks.
Accounting programs look boring. I’ve looked at accounting course descriptions. Lots of memorizing tax code and GAAP standards. Very little “build systems that generate insights.”
The tools look dated. QuickBooks Desktop looks like software from 2005 (because it is). Sage looks even worse. These don’t attract people who are used to working with modern developer tools.

Beancount is the first accounting tool that looked interesting to me as a technical person:

Plain text (I can diff it, version control it, grep it)
Programmable (Python ecosystem, I can build whatever I want)
Transparent (I can see exactly what’s happening, no black-box calculations)
Portable (it’s just text files, I can move them anywhere)

My “Controversial” Take

If accounting education doesn’t modernize, technical people will just bypass it entirely.

Right now I’m learning accounting through:

Beancount documentation
This forum
YouTube videos on double-entry bookkeeping
Reading IRS publications when I have tax questions
Trial and error with my own finances

Could I pass the CPA exam? No. Do I understand double-entry well enough to maintain books for small businesses? Yes. Am I going to spend 2 years in accounting school to get a credential that won’t increase my earning potential? Also no.

I think there are thousands of people like me—technical backgrounds, interested in finance, never considered accounting as a profession because the traditional path doesn’t appeal to us.

Plain text accounting might be the bridge.

If the profession wants to attract technical talent, you need to:

Make the tools interesting (Beancount does this)
Make the work intellectually challenging (segment 1 and 3 do this)
Make the compensation competitive (currently failing at this)
Remove unnecessary barriers to entry (the 150-hour requirement for CPA seems increasingly obsolete)

My Question for the Professionals Here

Do you want people like me entering the profession through non-traditional paths?

Or do you see self-taught technical bookkeepers as a threat to professional standards and client safety?

I genuinely don’t know the answer. Part of me thinks “accounting should have licensing requirements like engineering or medicine to protect the public.” Another part of me thinks “bookkeeping for a $500K/year coffee shop doesn’t need the same rigor as auditing a public company.”

Where should people like me fit in the three segments Mike described?