A Fourteen-Month-Old Tax Application Finally Landed - and Explains This Quarter's Entire Earnings Beat
Every quarterly filing since Q1 2014 has carried the same unresolved thread: MercadoLibre's Argentine subsidiary applied in May 2014 for a new software-development tax holiday, replacing the one that expired September 17, 2014, and the application sat pending for over a year while Argentina's effective tax rate ran at the full statutory level with no relief. On September 17, 2015 - the last day of this quarter - the Argentine Industry Secretary issued Resolution 1041/2015, approving the application and granting the holiday retroactively to September 18, 2014. MercadoLibre used this quarter's filing to catch up over a full year of foregone tax relief at once: a one-time income tax gain of approximately $16 million (of which $5.6 million relates to Q3 2015 itself and $3.2 million to Q4 2014, now recognized late) and a one-time labor-cost gain of approximately $4.2 million (60% relief on income tax and 70% relief on payroll taxes tied to software-development activities), partially offset by a $1.4 million reserve for law-related audit fees.
That single item is doing essentially all of the work behind this quarter's headline: net income attributable to MercadoLibre grew 35.2% to $45.6 million, and blended/effective tax rates cratered to just 4.2%/4.0% from Q3 2014's 34.9%/37.1%. But income from operations - the number the tax holiday doesn't touch - actually fell 4.0% to $45.3 million, with margin compressing to 26.9% from 31.9% a year ago, a 5.0-percentage-point decline on 14.0% reported revenue growth. Every dollar of this quarter's earnings growth, and then some, came from a one-time tax catch-up payment for a regulatory approval that took fourteen months to arrive - not from the business getting better. Read in isolation, "net income up 35%" is the kind of headline number this coverage exists to unpack; read against the operating-income line sitting right next to it, the real story is a segment-wide margin compression (see Five Country Segments below) that the tax windfall happens to be covering up almost exactly.
Reported net revenue grew 14.0% to $168.6 million, a genuinely narrower currency gap than recent quarters - local-currency growth was 67.9% (a 53.9-point gap, versus Q2 2015's 71.3 points) - but that narrower consolidated gap masks a new, and much larger, gap opening up specifically in Brazil: reported Brazilian revenue fell 5.0% even as it grew 47.3% in local currency, a 52.3-point gap and the first time this coverage has measured a reported YoY revenue decline in MercadoLibre's largest segment. The explanation is macro, not company-specific - Brazil's political crisis and fiscal deterioration through 2015 culminated in Standard & Poor's stripping the country of its investment-grade credit rating on September 9, 2015, and the real weakened sharply against the dollar through the quarter, translating directly into MercadoLibre's dollar-denominated Brazilian revenue line even as the underlying Brazilian business kept growing at a healthy clip in its own currency.
The Prescription
MercadoLibre should start disclosing operating income both with and without one-time discrete tax items in the same MD&A table it already uses for Adjusted Net Income, rather than leaving a reader to manually notice that a 35.2% net-income headline is sitting on top of a 4.0% operating-income decline. The Company already does this well for Venezuela-related charges (see the Adjusted Net Income reconciliation used in every prior quarter this coverage has covered); a one-time, fourteen-month retroactive tax catch-up of this size deserves the identical treatment, not a single paragraph buried in the effective-tax-rate footnote.
What it should stop doing: letting every segment's margin erode simultaneously without a consolidated explanation beyond individual country narratives. Brazil, Argentina, Mexico, Venezuela, and Other Countries all posted lower direct-contribution margins this quarter than a year ago - Mexico's fell 19.0 percentage points alone - and the filing addresses each in isolation rather than acknowledging the pattern as a single, coverage-wide margin story worth its own explanation. A reader has to build the segment table from scratch (see below) to even notice all five moved the same direction at once.
Key Financial Metrics
Three months ended September 30, 2015 vs. three months ended September 30, 2014 - consolidated, reported in USD (MercadoLibre reports natively in dollars; no FX conversion is needed)
| Metric | Q3 2015 | Q3 2014 | YoY |
|---|---|---|---|
| Net revenues | $168.6M | $147.9M | ✅ +14.0% reported (✅ +67.9% at local currency - a 53.9pp gap, narrower than Q2 2015's 71.3pp) |
| Cost of net revenues | $(56.8)M | $(43.4)M | ⚠️ +30.9% |
| Gross profit (66.3% margin) | $111.8M | $104.5M (70.7% margin) | ⚠️ +7.0%, margin -4.4pp |
| Product and technology development | $(17.0)M | $(13.6)M | ⚠️ +25.5% - mainly communication/maintenance and depreciation, not LTRP this quarter |
| Sales and marketing | $(31.1)M | $(29.4)M | ✅ +5.8% - well slower than revenue growth |
| General and administrative | $(18.4)M | $(14.4)M | ⚠️ +27.6% - roughly half of the increase is a base effect (Q3 2014 included a one-time tax-credit recovery gain that flattered last year's figure) |
| Income from operations (26.9% margin) | $45.3M | $47.1M (31.9% margin) | 🔴 -4.0%, margin -5.0pp - the real story this quarter |
| Foreign currency gain | $2.6M | $5.2M | ⚠️ Smaller currency gain than a year ago; no Venezuela devaluation event this quarter |
| Blended tax rate | 4.2% | 34.9% | 🔴 Almost entirely the one-time $16M Argentina tax-holiday catch-up, not a durable rate |
| Effective tax rate | 4.0% | 37.1% | Same driver |
| Net income attributable to MELI | $45.6M | $33.8M | +35.2% headline, but ⚠️ this is a one-time-tax-driven number, not organic growth |
| Diluted EPS | $1.03 | $0.76 | Same caveat applies |
| Free cash flow» | ~$1.2M | $42.1M | 🔴 -97% - a new office-related property advance and continued elevated capex, not a one-time acquisition this time |
| Capital expenditures (intangibles + property/equipment + PP&E advances) | $13.2M | $9.8M | ⚠️ +35% |
| Cash, ST & LT investments (period end) | $526.8M | n/a | Down 3.7% from Q2 2015's $546.9M |
| Loans payable and other financial liabilities (period end) | $297.0M | n/a | Almost entirely the $330M convertible notes, up as the debt discount amortizes |
Net income grew 35.2%, but income from operations - the line the Argentine tax holiday doesn't touch - fell 4.0% with margin down 5.0 percentage points. The entire earnings headline this quarter is a one-time tax-and-labor-cost catch-up payment for a regulatory approval fourteen months in the making, not evidence the operating business improved.
Key Operational Metrics
Three months ended September 30, 2015 vs. three months ended September 30, 2014
| Metric | Q3 2015 | Q3 2014 | YoY |
|---|---|---|---|
| Gross merchandise volume» (GMV, excludes motor vehicles/vessels/aircraft/real estate) | $1,842.1M | $1,682.7M | ✅ +9.5% - the first positive GMV growth this coverage has measured since Q4 2013 |
| Total payment volume» (TPV) | $1,384.4M | $975.1M | ✅ +42.0% |
| Total payment transactions | 22.0M | 12.5M | ✅ +76.0% |
| Confirmed registered users (cumulative, period end) | 138.4M | 115.2M | ✅ +20.1% |
| New confirmed registered users (in period) | 6.1M | 5.6M | ✅ +8.9% - recovered from Q2 2015's first-ever YoY decline |
| Items sold | 34.0M | 26.9M | ✅ +26.4% |
| Take rate» (net revenues / GMV) | 9.16% | 8.79% | ✅ +0.36pp |
| TPV as % of GMV | 75.2% | 58.0% | ✅ +17.2pp |
GMV growing 9.5% is the first positive reading this coverage has measured since Q4 2013 - a genuine reversal after five straight quarters of decline driven by currency devaluation outpacing real transaction growth. New registered user growth (+8.9%) also recovered from last quarter's first-ever YoY decline. Both are the clearest operational green shoots in this filing, and neither shows up in the profit lines above, which are dominated by the tax item and the segment-margin compression discussed next.
Five Country Segments - Every Segment Posted a Lower Margin, and Brazil's Reported Revenue Fell for the First Time in This Coverage
MercadoLibre reports the same five geographic segments adopted in Q3 2010 - Brazil, Argentina, Mexico, Venezuela, and Other Countries.
| Segment | Q3 2015 Revenue | YoY Revenue | Direct contribution margin, Q3 2015 | Direct contribution margin, Q3 2014 | Margin change |
|---|---|---|---|---|---|
| Brazil | $74.3M | 🔴 -5.0% | 41.1% | 44.7% | 🔴 -3.6pp |
| Argentina | $67.2M | ✅ +64.2% | 42.4% | 50.2% | 🔴 -7.8pp |
| Mexico | $9.9M | ⚠️ +0.3% | 13.4% | 32.3% | 🔴 -19.0pp |
| Venezuela | $8.9M | ⚠️ -4.5% | 58.6% | 71.7% | 🔴 -13.1pp |
| Other Countries | $8.4M | 🔴 -13.4% | 36.2% | 41.7% | 🔴 -5.5pp |
| Total | $168.6M | ✅ +14.0% | 40.7% | 46.9% | 🔴 -6.2pp |
Every single segment posted a lower direct-contribution margin than a year ago - the first time this coverage has measured a uniform margin decline across all five simultaneously. Mexico's collapse is the standout: revenue was essentially flat (+0.3%) while margin fell 19.0 percentage points to just 13.4%, the segment's worst quarterly reading this coverage has tracked. Argentina, despite growing revenue 64.2% and now representing 39.9% of consolidated revenue (up from 27.7% a year ago), still lost 7.8 points of margin - suggesting the growth is coming at a real cost-per-dollar-of-revenue price, not just currency mechanics. Brazil's -5.0% reported revenue decline (against +47.3% local-currency growth) is the headline figure, but its direct-contribution margin also fell independently of the currency translation, meaning the underlying Brazilian business itself is running less profitably, not merely reporting smaller dollar figures. Venezuela's ex-impairment margin kept falling for a second straight quarter (-13.1pp, following Q2 2015's -16.7pp real deterioration), even with no fresh devaluation event, on a segment now representing just 5.3% of revenue.
Beyond the Usual
The Argentine tax holiday's retroactive approval is a one-time item large enough to fully explain this quarter's net-income growth
Argentina's new software-development law tax holiday - applied for in May 2014, after the prior holiday expired September 17, 2014 - was finally approved by Resolution 1041/2015 on September 17, 2015, retroactive to the day the old holiday lapsed. MercadoLibre recorded the full twelve-month catch-up this quarter: approximately $16 million of one-time income tax gain (of which $3.2 million belongs to Q4 2014 and is only now being recognized) and $4.2 million of one-time labor-cost gain, less a $1.4 million audit-fee reserve. The Company states the aggregate per-share effect of the tax holiday was $0.15 for the quarter and $0.43 for the nine months - on a stock trading around $91, these are exact-dollar disclosures a reader can check directly against the $45.6 million headline net income figure.
This is a genuine, disclosed regulatory approval, not an accounting maneuver - but it means this quarter's entire earnings beat is a one-time, backward-looking item rather than a signal about current operating momentum. A reader who only saw "net income up 35.2%" without reading the tax-rate footnote would draw exactly the wrong conclusion about how the underlying business performed this quarter, which - on the operating-income line the tax holiday doesn't touch - actually went backward.
Free cash flow collapsed to essentially zero, and a new "advance for property and equipment" line is the largest single driver
Free cash flow» fell to roughly $1.2 million this quarter, down from $42.1 million a year ago and following Q2 2015's negative reading. Unlike Q2's swing (fully explained by two disclosed acquisitions), this quarter's weakness is driven mostly by a new $10.3 million "advance for property and equipment" cash outflow the filing doesn't further explain, on top of continued elevated capital expenditures ($13.2 million total, still above the $9.8 million spent a year ago). Operating cash flow itself remained solidly positive.
This is now two consecutive quarters of unusually weak free cash flow, each with a different one-time explanation (acquisitions in Q2, an unexplained property advance in Q3). Two different one-time causes in a row is itself worth tracking - if a third quarter posts a weak reading, "one-time" stops being a credible characterization.
The Brazilian real's collapse, not a change in the underlying business, explains Brazil's first-ever reported revenue decline
Brazil's reported revenue fell 5.0% even as it grew 47.3% in local currency - a 52.3-percentage-point currency gap, the widest this coverage has measured for any single segment in any quarter, wider even than Venezuela's currency-devaluation gaps at their peak. Standard & Poor's downgraded Brazil's sovereign credit rating below investment grade on September 9, 2015, amid a deepening political and fiscal crisis, and the Brazilian real weakened sharply against the dollar through the quarter as a direct consequence.
This underlying-business-strong, reported-numbers-weak split is exactly the same currency mechanic this coverage has tracked in Venezuela and Argentina for two years - now showing up, for the first time at this scale, in MercadoLibre's largest and previously most currency-stable segment.
The Rio de Janeiro customer-service claim and an Argentine labor-union claim both closed this quarter
The Rio de Janeiro customer-service claim - already effectively resolved in Q2 2015 when both the Superior Court of Justice and Federal Supreme Court declined to hear the state prosecutor's appeal - remains closed. Separately, a previously undisclosed Argentine labor-union claim (filed September 29, 2014 by the Federación Argentina de Empleados de Comercio y Servicios, seeking roughly $0.43 million in allegedly unpaid employee benefits over ten years) was settled for ARS 2.4 million plus interest and legal expenses on August 12, 2015, and the case is now closed - a small, immaterial resolution but a new claim this coverage hadn't previously tracked.
The São Paulo tax claim remains active judicial litigation with zero reserve for a ninth consecutive filing
The lawsuit before the 9th Treasury Court of the City of São Paulo - contesting taxes and fines assessed against MercadoLibre's Brazilian subsidiary for the 2007-2010 period - remains exactly where it has sat since November 2013: the lower court's ruling still pending. The R$41 million court deposit is now valued at $10.3 million, down from $13.2 million at June 30, 2015 - purely the same Brazilian real devaluation discussed above shrinking the deposit's dollar value, not any partial repayment.
This is the ninth consecutive filing this coverage has flagged the identical pattern: an eight-figure court deposit sitting in escrow, active judicial litigation approaching its second year with no ruling, and a "remote" risk classification unchanged since October 2013. The separate, smaller 2005-2007 São Paulo tax claim sits in the identical unresolved posture.
Coverage Table
| Theme | Q3 2015 | Q3 2014 | YoY | Why it matters |
|---|---|---|---|---|
| Net revenue, reported (USD) | $168.6M | $147.9M | ✅ +14.0% | Local-currency growth of 67.9% shows most of it is currency-masked, though the consolidated gap narrowed |
| Income from operations | $45.3M (26.9% margin) | $47.1M (31.9% margin) | 🔴 -4.0%, -5.0pp | The real story this quarter - the tax holiday doesn't touch this line |
| Net income attributable to MELI | $45.6M | $33.8M | +35.2% headline | Almost entirely a one-time ~$20.2M Argentina tax/labor catch-up, not organic growth |
| GMV | $1,842.1M | $1,682.7M | ✅ +9.5% | First positive GMV growth since Q4 2013 |
| Free cash flow | ~$1.2M | $42.1M | 🔴 -97% | Second straight quarter of unusually weak FCF, different one-time cause each time |
| Every segment's direct contribution margin | Lower than Q3 2014 in all five segments | n/a | 🔴 | First uniform margin decline across all segments this coverage has measured |
| São Paulo tax claim | Same posture as Q2 2015, lower court ruling still pending | n/a | 🔴 | Ninth consecutive filing flagged; still "remote," zero reserve |
Target Valuation Range
DCF fair-value range: roughly $2,943 million (base case) to $6,101 million (bull case) enterprise value, against a $3,790.4 million actual enterprise value - the stock now sits just above the base case, inside the fair-value band for the first time in this coverage. The stock fell 35.7% this quarter amid Brazil's sovereign downgrade and a broader emerging-market selloff, closing most of the valuation gap this coverage flagged reopening last quarter - not because the business got cheaper on its own merits, but because the price did.
The stock fell 35.7% this quarter to $91.06, down from June 2015's $141.70 (MercadoLibre has never split its common stock since its August 2007 IPO, so no split adjustment applies). Netting cash against loans payable brings enterprise value down 36.8% from Q2 2015's $5,998.0 million.
| Market cap → enterprise value | Q3 2015 |
|---|---|
| Share price (period-end) | $91.06 |
| Shares outstanding | 44,156,854 |
| Market capitalization | $4,020.2 million |
| Plus: loans payable | $297.0 million |
| Less: cash & investments | $526.8 million |
| Enterprise value | $3,790.4 million |
| Valuation multiples | Q2 2015 | Q3 2015 | Change |
|---|---|---|---|
| TTM Net Revenue | n/a | $632.4 million | - |
| Enterprise value | $5,998.0 million | $3,790.4 million | ⬇ down |
| EV/Sales | 9.81x | 5.99x | ⬇ down |
| P/E (TTM) | 70.2x | 39.8x | ⬇ down (tax windfall inflates net income) |
| EV/FCF (TTM) | 55.4x | 56.3x | ~flat |
Trailing free cash flow of roughly $67.3 million is down from Q2's $108.2 million, on the weak Q3 FCF discussed above.
| Scenario | Key assumption | Implied EV | % of actual EV |
|---|---|---|---|
| Current (Q3 2015 close) | actual market price | $3,790.4 million | 100% |
| Base | Revenue growth 20%→10% over 5yrs; FCF margin 10.6%→28%; 13% discount, 4% terminal growth | $2,943 million | 77.6% |
| Bull | Revenue growth 24/20/17/14/11%; FCF margin 26%→34%; 11% discount, 5% terminal growth | $6,101 million | 161.0% |
The base case's coverage jumped from Q2 2015's ~48.9% to 77.6% - not because the underlying cash-flow trajectory improved, but because the stock cratered faster than the DCF's own fair-value estimate did. The bull case is a full reversal of Q2 2015's first-ever miss since Q3 2014 (98.4%), now clearing the actual price by 61%.
Reverse DCF: holding the 13% discount rate and solving for the perpetual FCF growth rate that would justify today's $3,790.4 million enterprise value on the $67.3 million trailing FCF base gives roughly 11.0% growth, forever - essentially unchanged from Q2 2015's ~11.0%, since enterprise value and trailing free cash flow both fell by nearly the same proportion this quarter, leaving the market's implied growth expectation about where it was.
The valuation gap closed dramatically this quarter, but almost entirely on the price side of the equation - the stock crashed, not the DCF's fair-value estimate rising to meet it. Whether the underlying operating-margin compression flagged across every segment this quarter proves temporary (a Brazil-specific currency and macro shock) or the start of a durable trend, and whether free cash flow recovers from two consecutive weak quarters, are the clearest swing factors heading into Q4.
MercadoLibre, Inc.'s Quarterly Report on Form 10-Q for the quarter ended September 30, 2015, filed with the U.S. Securities and Exchange Commission and signed November 5, 2015. Historical MELI share price data covers month-end closes from October 2013 through September 2015; MercadoLibre has never split its common stock since its August 2007 IPO, so no split adjustment applies to these figures. No presentation, press release, or transcript was located for this filing, so this post does not include a management-commentary section.