Weekly Signals Desk | Five Dividend Increases Flagged by the FMP API (Week of Sept 7- 11)

The signal desk run over the past two weeks surfaced five declarations separated by an order of magnitude in size: Argan lifted its quarterly payout by 40%, while U.S. Bancorp moved 3.8%. Stewart Information Services, Altria and Nordson filled the space between, each for a different reason.

That spread is the story. A dividend increase is one of the few capital-allocation decisions a board has to put a precise number on, and the number carries information only when it is read against the base it came from. Using the FMP Dividends Calendar API, this article works through what these five declarations indicate about cash generation, payout capacity and management priorities, and how the same endpoint can be used to build a repeatable screen rather than a one-off list.

Key Takeaways

• The percentage increase is the least informative number in a dividend declaration. Argan's 40% raise consumes a fraction of the cash that Altria's 4.7% adjustment does, because the starting bases differ by more than a factor of two on an annualized per-share basis.
• Three distinct funding logics show up in the same two-week window: balance-sheet capacity at Argan, regulatory capital headroom at U.S. Bancorp, and a multi-decade compounding record at Nordson supported by backlog growth.
• Yield and raise size move in opposite directions across this group. The two lowest-yielding names posted the two largest increases, which reframes those declarations as capital-allocation statements rather than income events.
• The Dividends Calendar API establishes the event; cash flow, segment revenue and ratio datasets determine whether the new payout has support behind it.

Five Payout Increases and What Sits Behind Them

Argan (NYSE: AGX)

Argan declared a quarterly dividend of $0.70 per share, or $2.80 annualized, a 40% increase over the prior $0.50 payment. The dividend is payable on October 30, 2026, to stockholders of record on October 22, with an ex-dividend date of October 21. The annual yield is 0.7%.

This is the largest percentage move in the group and the smallest yield, which is usually a sign that the payout is not the point. Argan's most recent quarter showed revenue up roughly 60% year over year to $384 million, with adjusted EBITDA margin expanding about two points and the power segment carrying most of the growth. The company holds over $1 billion in cash and investments against no debt. Against that position, a 40% raise is a modest deployment of capital, running alongside share repurchases rather than replacing them.

The detail worth tracking is the tension underneath the headline. Project backlog declined from roughly $2.9 billion to $2.5 billion over two quarters, which is what happens when work converts into revenue faster than new awards replace it. That is not a problem in itself, but it makes the pace of new bookings the variable that matters more than the payout. FMP's Cash Flow Statement API is the cleanest way to see whether the raise is being funded from operating cash generation or from the accumulated balance, since engineering and construction businesses can show large working-capital swings that make a single quarter's earnings a poor proxy for distributable cash.

U.S. Bancorp (NYSE: USB)

U.S. Bancorp declared a quarterly dividend of $0.54 per share, equal to $2.16 annualized. The increase is 3.8% from the prior $0.52 payment. It is payable on October 15, 2026, to shareholders of record on September 30, with an ex-dividend date of September 29. The annual yield is 3.5%.

Bank payouts are governed by a different constraint than industrial ones, and this declaration reflects that directly. The company's stress capital buffer was set at 2.6% and holds through October 2027, putting its common equity Tier 1 floor at 7.1% against a reported ratio near 10.8%. The headroom is substantial, yet the dividend moved less than four percent. The reason is visible in the other half of the capital plan: roughly $4.1 billion remained available under an existing $5 billion repurchase authorization. Management is choosing the more flexible lever, and a single-digit dividend raise alongside a large buyback capacity is a standard signal of that preference.

For a bank, the payout question is inseparable from asset quality and funding mix, so the follow-up work belongs in the balance sheet rather than the income statement. FMP's Balance Sheet Statement API allows deposit composition, loan balances and equity to be tracked across quarters, which is where pressure on a distribution would appear first. Credit provisioning and net interest margin trends would complete that picture, and neither is reliably inferred from a dividend declaration on its own.

Stewart Information Services (NYSE: STC)

Stewart Information Services declared a quarterly dividend of $0.55 per share, or $2.20 annualized, a 4.8% increase from the prior $0.525 payment. The dividend is payable on September 30, 2026, to shareholders of record on September 15, with an ex-dividend date of September 14. The annual yield is 3.1%.

A mid-single-digit raise at a title insurer reads as measured given what the top line did. Second-quarter revenue rose roughly a quarter year over year to just under $900 million, but the composition of that growth deserves attention. Agency title revenue grew far faster than direct title revenue, and the real estate solutions segment expanded sharply with help from an acquisition. Agency business carries materially lower retained economics than direct operations, and acquired service revenue arrives with integration costs attached. Operating expenses climbed as a share of revenue by nearly three points over the same period.

That mix shift is why the raise looks appropriately sized rather than conservative. Revenue growth of this shape does not convert to distributable cash at the same rate that direct title growth would, and title volumes remain tied to a transaction cycle the company does not control. The title loss ratio improving to roughly 3.2% is the encouraging counterweight, since claims experience is the main way underwriting surprises reach the dividend. To separate genuine operating leverage from mix effects here, FMP's Revenue Product Segmentation API is the relevant dataset, because consolidated revenue growth at this company can conceal more than it reveals.

Altria Group (NYSE: MO)

Altria declared a quarterly dividend of $1.11 per share, equivalent to $4.44 annualized, up 4.7% from $1.06. The payment is due on October 9, 2026, to shareholders of record on September 15, with an ex-dividend date of September 14. The annual yield is 6.6%.

In cash terms this is the heaviest commitment on the list, and it arrives against a declining volume base. Cigarette shipment volumes continued to fall in the most recent quarter, and the company's oral nicotine brand lost volume in a category that is otherwise expanding. Neither of those trends is new, and neither prevented the board from extending an increase record that now spans more than five decades. What funds the raise is arithmetic rather than growth: pricing absorbs much of the volume decline, and a share count reduced by roughly a seventh over ten years means total cash paid out rises far more slowly than the per-share figure suggests.

That mechanism is durable but not unlimited, which is the analytical point. A 6.6% yield paired with a mid-single-digit raise implies the market is already pricing a slow erosion of the underlying business, so the dividend signal here is less about confidence and more about the pace at which pricing and buybacks can continue offsetting declining units. FMP's Key Metrics TTM API puts free cash flow per share and payout coverage on a rolling basis, which is the right frame for a payer whose per-share economics are being managed as deliberately as its operations.

Nordson Corporation (NASDAQ: NDSN)

Nordson declared a quarterly dividend of $0.94 per share, or $3.76 annualized, a 14.6% increase from the prior $0.82. It is payable on October 2, 2026, to holders of record on September 10, with an ex-dividend date of September 9. The annual yield is 1.1%.

This is the second-largest raise in the group and it comes from a company with the longest uninterrupted record, now past six consecutive decades of annual increases. That combination is unusual. Long-streak payers typically protect the streak with small, safe increments, and a double-digit step signals that the board is comfortable resetting the base rather than merely defending continuity. The most recent quarter supports that reading: sales grew around 10% to $818 million with EBITDA margin holding near 32%, and the advanced technology segment grew close to 30%.

The number that carries the most weight is order backlog, which rose roughly 35% year over year. For a precision dispensing and test business, backlog is the closest available read on the next several quarters of revenue, and a raise of this size following that kind of order growth is consistent with management treating the demand as durable rather than a single strong period. Free cash flow through three quarters was up over 10% against the prior year, which is what makes the step affordable. The full history of the streak, including how the size of each annual increment has changed across cycles, is retrievable through FMP's Dividends Company API, and that sequence is more informative than any single declaration.

Reading the Spread: Why Raise Size and Signal Strength Diverge

Line these five up by percentage increase and the ordering is almost exactly inverted against yield. The two largest raises came from the two lowest-yielding names, and the smallest raise came from a company with abundant capital headroom. That is not a coincidence, and it is the most useful pattern in the group. Where a dividend is a small share of cash generation, the board can move it sharply without committing much. Where the dividend already consumes most of distributable cash, the increase becomes an exercise in precision, and a 4.7% step at Altria represents a far larger absolute commitment than a 40% step at Argan.

The second pattern is that the constraint differs by business. Argan's limit is how quickly new project awards replace converted backlog. U.S. Bancorp's is regulatory, and its dividend decision is effectively a residual after buyback preference and buffer requirements are set. Stewart's is cyclical and mix-driven. Altria's is volumetric. Nordson's is the least binding of the five, which is why it produced the most confident move. Grouping these declarations as a single “dividend increase” theme flattens exactly the information that makes them worth reading.

Working across the FMP endpoint set is what allows that separation to be made systematically rather than case by case. Coverage is the first test, and the Cash Flow Statement API answers it by placing dividends paid against operating cash flow and capital expenditure in the same series. The Financial Ratios TTM API then normalizes payout ratios and free cash flow yields across companies whose income statements are not directly comparable, which matters when a bank, an insurer and two industrials appear in the same screen. Where leverage is part of the funding question, the Enterprise Values API provides the net debt context that a payout ratio alone omits.

The final layer is whether expectations agree with the decision. The Financial Estimates API shows where consensus forecasts sit for revenue and earnings over the coming periods, and comparing that trajectory against the size of a raise is often more revealing than either figure alone. A board increasing its distribution into rising estimates is making a different statement than one increasing into flat or falling forecasts. Where those two series point the same direction, the declaration reinforces what the fundamentals already show. Where they diverge, the dividend becomes a reason to look harder at the gap rather than a conclusion in itself.

From Declaration to Insight: Building a Repeatable Dividend Screen via FMP API

If dividend adjustments are going to function as usable signals, the process has to begin at the point where the decision actually occurs: the declaration itself. That means sourcing the data directly from the FMP Dividends Calendar API, which captures dividend announcements at the moment companies publish them, before those entries are absorbed into broader aggregated datasets.

Before running any queries, confirm that your API key is active. Once authenticated, the Dividends Calendar endpoint effectively becomes the intake layer for the entire workflow. It returns a structured dataset containing the ticker symbol, declared dividend amount, key payout dates (declaration, record, payment, and ex-dividend), yield, and payment frequency. That initial pull forms the starting universe from which dividend changes can be identified and analyzed.

Endpoint:

https://financialmodelingprep.com/stable/dividends-calendar?apikey=YOUR_API_KEY

Sample Response:

[ { "symbol": "1D0.SI", "date": "2025-02-04", "recordDate": "", "paymentDate": "", "declarationDate": "", "adjDividend": 0.01, "dividend": 0.01, "yield": 6.25, "frequency": "Semi-Annual" } ]

Step 1: Capture Recent Declarations

Start by querying the Dividends Calendar over a short, controlled time frame—typically the most recent 10 to 14 days. This window is long enough to capture new declarations while limiting contamination from older entries that sometimes reappear due to reporting delays. The output from this step forms the working universe for the rest of the analysis.

Step 2: Stack It Against the Prior Dividend

Next, for every ticker surfaced in the initial pull, retrieve the previous dividend using the historical dividend endpoint. This historical anchor is critical. Without it, unchanged recurring payments and true increases are indistinguishable. The comparison introduces context and allows the workflow to focus on intent rather than repetition.

Step 3: Filter for Material Moves

With both the new and prior dividend values in hand, calculate the percentage change using

(New Dividend − Old Dividend) ÷ Old Dividend × 100.

Apply your screening criteria to narrow the list. A common approach is to flag increases of 5% or more paired with an annual yield of at least 2%, which helps remove token raises while preserving economically relevant moves. Thresholds can be tuned depending on whether the focus is income generation, payout discipline, or signal detection.

Example Workflow: Detecting 5%+ Dividend Hikes

1. Pull a fresh 14-day window from the Dividends Calendar API.
2. For each ticker, fetch its prior payout via the historical dividend endpoint.
3. Compute the percentage change using the formula above.
4. Keep only companies posting 5%+ increases and yielding 2% or more.

Expanding Your Dividend Tracking Setup

Most dividend screens start out as simple monitoring systems. The initial goal is usually straightforward: capture new declarations quickly enough to feed a watchlist, weekly report, or alert workflow before the information gets absorbed into broader market data. Using the Financial Modeling Prep Free plan, that process stays lean and reactive, centered on pulling fresh entries from the Dividends Calendar as they are published.

The analysis becomes more useful once historical comparison enters the workflow. Access to roughly one year of dividend history through the Starter plan makes it possible to place each declaration against its prior payout rather than treating every entry as a standalone event. That shift matters because recurring dividends and genuine increases often look identical in raw calendar data. Once the historical layer is added, patterns begin to emerge around consistency, timing, and how management teams behave across different operating environments.

A longer historical window changes the screen again. With up to five years of dividend history available through the Premium, payout decisions can be evaluated across multiple business cycles instead of isolated reporting periods. At that depth, dividend changes stop functioning as simple event flags and start becoming part of a broader operating history — one that can be compared against earnings pressure, margin expansion, sector slowdowns, or shifts in capital allocation strategy over time.

When a Desk Tool Turns Into Firmwide Infrastructure

Most market-monitoring workflows do not begin as institutional systems. They start as tightly scoped analyst processes — a dividend screen running weekly, a historical comparison model maintained on a single desk, or a signal tracker built to support a specific coverage universe. The transition happens when those outputs start influencing conversations outside their original context. Once portfolio managers, strategy teams, and risk committees begin referencing the same screen, consistency becomes more important than the screen itself.

At that point, the analyst who built the workflow often becomes something else entirely: the internal advocate for standardization. The challenge shifts away from finding signals and toward ensuring every team is interpreting the same dataset under the same assumptions. That means aligning declaration feeds, refresh schedules, historical comparison windows, and filtering logic across desks. Without that coordination layer, firms gradually accumulate fragmented versions of the same model — similar screens producing slightly different conclusions because the underlying methodology drifted over time.

Centralized workflows reduce that friction. When dividend declaration data, historical payout comparisons, cash flow metrics, and analyst estimate revisions feed into a shared dashboard environment, research teams can spend less time reconciling discrepancies and more time debating interpretation. A portfolio manager reviewing a dividend signal should be looking at the same benchmark logic as the analyst who generated it and the risk team evaluating its exposure implications. Shared infrastructure turns isolated analysis into an institutional reference point.

As adoption broadens, governance naturally becomes part of the analytical process. Questions around lineage, revision tracking, access controls, and auditability start carrying as much weight as the signal itself. Teams want to know whether historical dividend records were revised, whether screening thresholds changed between quarters, and whether prior outputs can be reproduced under the same assumptions. In institutional environments, repeatability is credibility.

That is usually the stage where successful desk-level workflows migrate onto more formal infrastructure layers, not to change the analytical framework, but to preserve it as usage expands across teams. In practice, that often means consolidating data access and workflow management through systems designed for broader internal distribution, such as the Financial Modeling Prep Enterprise Plan, where the emphasis shifts toward stability, consistency, and operational transparency across the research stack.

What a Fortnight of Declarations Adds Up To

Stretching the window to two weeks made the divergence legible in a way a single week rarely does: five boards, five different constraints, and almost no relationship between the size of the raise and the weight of the decision behind it. Read that way, and paired with the coverage data that sits one query away, the FMP Dividends Calendar API turns a routine stream of declarations into a running record of how management teams are choosing to deploy cash.

If you found this useful, you might also like: Signals Desk Weekly Take via FMP API | Five Companies With Persistent Earnings Beats (Aug 31-Sept 4)

Disclosure: Signals Desk content is provided for informational and analytical purposes only and does not constitute investment advice or trade recommendations. The analysis reflects interpretation of market data and publicly disclosed or third-party information, including data accessed via Financial Modeling Prep APIs, at the time of publication. Signals discussed are probabilistic, can be wrong, and may change as market conditions and consensus data evolve. This content should be considered alongside broader research, individual objectives, and risk assessment.