Switching Jobs Mid-Year: Why Two Employers Under-Deduct Tax
In this article (5 sections)
Related programme: the Data Science course (6 months; ₹1,30,000 (inclusive of GST) · EMI from ₹5,000/month).
When you change jobs partway through a tax year, each employer usually deducts TDS only on the salary it pays you. Each also allows the full standard deduction and, where income is low enough, the full rebate. Your tax, though, is charged on the combined salary, with one standard deduction. In the hypothetical tax year 2026-27 switch below, the two employers deduct ₹44,070 against a liability of ₹1,16,220, leaving ₹72,150 to pay when you file. Declaring the first salary to the second employer spreads that sum across your remaining payslips instead.
The switch
| Employer A | Employer B | |
|---|---|---|
| Months | April–September 2026 | October 2026–March 2027 |
| Monthly gross | ₹1,20,000 | ₹1,50,000 |
| Salary paid | ₹7,20,000 | ₹9,00,000 |
All figures are hypothetical, on the default new regime. Combined salary for tax year 2026-27 (1 April 2026 to 31 March 2027) is ₹16,20,000.
What each employer sees
Section 392(1) of the Income-tax Act, 2025_Bill.pdf) requires TDS at the average rate on the employer's estimate of your salary for the tax year.
Employer A assumed you would stay all year and estimated ₹14,40,000. Taxable income after the ₹75,000 standard deduction was ₹13,65,000, giving annual tax of ₹88,140 with cess. It deducted one twelfth each month, ₹7,345, and ₹44,070 by the time you left.
Employer B knows nothing of A's salary. It estimates ₹9,00,000 for its six months. Less the standard deduction, that is ₹8,25,000, within the section 156 rebate. It deducts nothing.
| Estimate used | Tax on that estimate | TDS actually deducted | |
|---|---|---|---|
| Employer A | ₹14,40,000 | ₹88,140 | ₹44,070 (six months) |
| Employer B | ₹9,00,000 | ₹0 | ₹0 |
| You, on actual salary | ₹16,20,000 | ₹1,16,220 | ₹44,070 |
Your real taxable income is ₹16,20,000 − ₹75,000 = ₹15,45,000. Tax is ₹60,000 on the first ₹12,00,000 plus 15% of ₹3,45,000 (₹51,750), making ₹1,11,750, and ₹1,16,220 with 4% cess. The shortfall is ₹72,150.
Two things caused it. Employer B applied the rebate, which is lost once combined income passes ₹12,00,000. And A's deduction, though based on a full year, stopped after six months, so it covered only half the tax on its own estimate. The slab rates are listed on the department's slab page.
The fix: declare the earlier salary
Section 392(4) lets the employer, at your option, take into account salary you received from another employer in the same tax year. Give Employer B the salary A paid and the TDS it deducted, in the form B asks for. B then estimates your whole year:
| Step | Amount |
|---|---|
| Tax on combined salary | ₹1,16,220 |
| Less TDS already deducted by A | ₹44,070 |
| Left to deduct | ₹72,150 |
| Over six remaining months | ₹12,025 a month |
If you declare it late, the same ₹72,150 is spread over fewer months. Declared in January, with three payslips left, it becomes ₹24,050 a month. Section 392(5)(c) allows the employer to adjust for a shortfall within the year, but the adjustment cannot run past March.
# Hypothetical switch. New regime, tax year 2026-27.
def tax(ti):
t, lo = 0.0, 0
for hi, r in ((400000, 0), (800000, .05), (1200000, .10), (1600000, .15),
(2000000, .20), (2400000, .25), (float("inf"), .30)):
t += max(0, min(ti, hi) - lo) * r
lo = hi
return 0.0 if ti <= 1200000 else round(min(t, ti - 1200000) * 1.04)
SD = 75000
a_months, a_monthly = 6, 120000
b_months, b_monthly = 6, 150000
a_tds = tax(12 * a_monthly - SD) / 12 * a_months # A assumed a full year
b_tds_unaware = tax(b_months * b_monthly - SD) # B sees only its own pay
actual = tax(a_months * a_monthly + b_months * b_monthly - SD)
shortfall = actual - a_tds - b_tds_unaware
print("A deducted:", a_tds, "B deducted:", b_tds_unaware)
print("actual tax:", actual, "shortfall:", shortfall)
for months_left in (6, 3):
print(months_left, "months left ->", round(shortfall / months_left), "a month")If you file with a shortfall
If you did not declare the first salary, the ₹72,150 is still due. It is paid with your return, after checking both salary certificates against your payslips. Checking Form 16 Part B against your payslips shows the reconciliation for one employer; repeat it for each, then add the two. Budget for the amount from the day you join the second employer. It is easier to save ₹12,025 a month than to find ₹72,150 in July.
Variations that change the size of the gap
- Employer A trued up before you left. If A knew your exit date, its final payslip may have recomputed on actual salary only. On ₹7,20,000 that is within the rebate, so A might have deducted less, not more, and the shortfall would be larger.
- A gap between jobs. Fewer paid months lower the combined salary and the shortfall. The mechanics are the same.
- A joining bonus from B. That is salary too, and B may include it in its own estimate. Declare A's salary anyway, so that B deducts on the whole year.
For how an employer builds its monthly figure from the annual estimate, see how your employer calculates monthly TDS.
Combining two partial records into one correct total is a routine data-joining problem. It is practised in our Data Science programme, live online with optional in-person labs at our Noida centre.
Frequently asked questions
Is under-deduction by two employers a mistake by either of them?
Not necessarily. Section 392(1) asks each employer to deduct on its estimate of salary, and an employer who has not been told about your earlier salary can only estimate its own. The shortfall comes from the missing information, which is why declaring the earlier salary is the fix.
Pankit Kumar has 10 years in Data Science & AI, building and shipping production systems in regulated pharma and clinical environments. He is a freelance trainer at Boston Institute of Analytics, AnalytixLabs and Scaler, and has taught this material to thousands of working professionals.
This article is part of our Data Science programme — 6 months. From data foundations to machine learning, deep learning and deployment.
Explore Data Science