Currency-market participants differ in funding, execution access, technology and objectives. A retail account and an institution are not uniform categories: costs and constraints vary within both. Compare the actual instrument, legal entity and execution terms before inferring an advantage from size alone.

Who is who, and who moves the volume

BIS turnover surveys classify reporting dealers, other financial institutions and non-financial customers. These are counterparty and transaction categories, not a ranking of trading skill or profit. They cannot be converted into a global retail-versus-institutional percentage without an explicit definition and dataset. Retail CFD turnover, account loss warnings, people and individual trades have different units and denominators; adding them does not produce one meaningful share. A dealer manages client-flow exposure, a portfolio manager pursues a mandate, and a corporate treasurer may hedge a payment rather than seek a standalone trading gain.

A large order may incur market impact, but impact depends on available liquidity, timing and execution method. Small orders can also suffer slippage or rejection. Size alone does not guarantee a price-neutral fill. The market-maker and interbank guides explain the relationships. FX is fragmented: co-location at one system is not a position beside the entire market. Under FX Global Code Principles 8–9, an agent executes under a client mandate without taking its market risk, while a principal trades on its own behalf. Wholesale good practice does not replace applicable law or a retail broker’s contract; a CFD client contract and the provider’s later hedge are separate relationships.

Four accounts are not four people

This fictional sample contains two people: A has three active CFD accounts and B has one. Over the same 12-month period, each account belonging to A has a net result of −PLN 10, while B has +PLN 30. All held CFD positions; all costs and the result on positions still open at the period end are included.

Scroll sideways to see the full diagram

Accounts and people have different denominatorsPerson A has three accounts with negative net results; person B has one with a positive result. Three of four accounts, or seventy-five percent, lose money. One of two people, or fifty percent, has at least one losing account. The second statistic is a separately defined measure of people, not a substitute for the account percentage.Same fictional sample; two countsPerson APerson BA1A2A3B1−PLN 10−PLN 10−PLN 10+PLN 30Losing accounts: 3 / 4 = 75%People with ≥1 losing account: 1 / 2Second measure: 50%; a different unit.
Both bars have the same total length: the first divides four accounts, the second two people. Their red portions count different units.

3 / 4 = 75% describes accounts. 1 / 2 = 50% describes people under our separate definition. The first person appears three times in the account numerator but once in the people numerator. The cited Polish KNF method counts CFD accounts; do not turn its percentage into the loss probability of a new client or an individual trade. A real dataset needs separate person identifiers and account aggregation rules.

Assumptions: Two fictional people and four active CFD accounts with one fictional provider, over one shared 12-month period. Net results include realised and unrealised P/L and all CFD costs; deposits, withdrawals and results from other products are excluded. Our measure of people means “has at least one losing account”. No current broker percentage or market-wide statistic is asserted. Official source: Polish KNF decision DAS.456.2.2019, point 7(f).

Leverage and the rules of the game differ by side

Retail CFD protections depend on jurisdiction, provider, product and legal client classification. ESMA’s temporary intervention, initially effective in August 2018, included a 30:1 major-currency-pair limit and lower limits for other categories. Those ESMA measures expired in July 2019; check the national measure applicable to the actual account rather than treating the old intervention as a permanent worldwide cap. Poland’s KNF decision defines a distinct experienced retail category affecting initial margin while retaining other retail protections. Its loss warning counts the provider’s CFD accounts over the preceding 12 months and is recalculated at least every three months. It is not the loss probability of every forex client or the next trade.

Institutional funding and leverage arrangements vary with the participant, instrument and applicable regulation. Professional classification does not remove prudential requirements, contractual limits or market risk. For a retail account, calculate exposure relative to equity and the actual margin requirement; a regulatory cap is not a guarantee against loss.

The asymmetry of information and technology

Some institutions invest in low-latency infrastructure and specialised research. Their access and processing speed differ; not every institution has co-location or identical terminals. Evaluate the information source and execution delay relevant to the strategy instead of assuming one universal institutional setup.

Retail data may be live or delayed depending on the feed and subscription. Internet latency and execution rules still matter for a strategy sensitive to milliseconds. Test the actual setup; neither live data nor a slower timeframe establishes profitability.

Compare the complete cost, not the spread alone

Compare spread, commission, financing, conversion and execution effects for the same instrument and size. The following fictional tariffs are not typical retail or wholesale offers. Assume EUR/USD midpoint prices of 1.1000 on entry and 1.1020 on exit, symmetric quotes, a constant spread in each tariff and full fills at the stated ask and bid. Buy then sell EUR 10,000 in a USD account. Exclude extra slippage, financing, fixed charges and conversion; the stated commission is the total for both sides.

Hypothetical round trip: a narrower spread can cost more
Tariff A: spread 3 pips, total commission USD 0Buy 1.10015; sell 1.10185
Tariff A net result(1.10185−1.10015) × 10,000 = USD 17
Tariff B: spread 1 pip, total commission USD 2.50Buy 1.10005; sell 1.10195
Tariff B net result(1.10195−1.10005) × 10,000−2.50 = USD 16.50
Cost versus the USD 20 midpoint moveA: USD 3; B: USD 3.50

The midpoint move is 20 pips, worth USD 20 for this size. Execution-price differences already include the spread: do not deduct it again. Tariff B has USD 19 before its USD 2.50 commission, leaving USD 16.50 versus A’s USD 17. Add the charges excluded here in a real comparison and check representative times, sizes and fills; multiplying this table by a thousand may change the applicable price and tariff. Reducing turnover may reduce some charges but also changes opportunities and exposure. Less frequent trading is not by itself evidence of a profitable edge.

Match the method to available resources

A home setup may be unsuitable for a latency-dependent strategy, and a small account limits the amount of capital available. Neither point means that every retail trader has delayed information or cannot diversify. Assess the particular resources, concentration and loss tolerance needed by the method.

Choose methods whose data, execution and capital requirements you can meet, then test their net results. Resources do not determine an account’s outcome in advance, and provider loss-account statistics do not identify the result of one individual strategy.

Freedom to wait is not proof of an edge

Small size may reduce market impact, but it does not remove slippage or execution risk. An individual trading personal funds may choose to remain flat. Institutional mandates vary and can also permit cash or no position; the distinction is not a universal inability to wait.

A journal and consistent rules make decisions easier to audit. They do not eliminate most errors automatically or demonstrate an edge. Record costs, deviations and realised outcomes, then test whether the method meets its stated objectives.

What to do if you trade as a retail trader

  1. Select a method whose data and execution requirements you can meet. Check the relevant timeframe, latency, capital and costs in testing; no timeframe guarantees an advantage from small size.
  2. If your plan permits remaining flat, use that choice when its entry conditions are absent. Check any mandate or contractual restrictions rather than assuming institutions always have to trade.
  3. Set position size from an explicit loss budget and include spread, commission and financing. Smaller or fewer trades can reduce some costs, but do not guarantee that an account survives or that its method improves.
  4. Record each decision, its reason, execution and full net result. Review deviations from your rules and the distribution of outcomes. A journal supports risk review; it does not by itself remove most causes of loss.

Institutional and retail participants have different, variable resources and constraints. Compare the actual conditions rather than assuming a predetermined winner. The COT report describes reportable futures and options positioning, not all global spot-FX orders. If considering managed accounts, assess their mandate, costs, custody and risks separately.

A standalone hedge loss can still meet a company’s stated objective. ITA distinguishes currency-risk management from pursuing trading profit. In our hypothetical Polish importer model, EUR 100,000 is due in 30 days and the payment budget is PLN 435,000. A deliverable forward fixes purchase of the full invoice amount at 4.3000 PLN/EUR, costing PLN 430,000. Assume matching amount and date, performance of the contract and no fees or financing costs.

Own importer model: same invoice, two hypothetical delivery prices
Spot 4.1000: purchase without hedgePLN 410,000; within budget
Spot 4.5000: purchase without hedgePLN 450,000; above budget
Forward in either scenarioPLN 430,000; within budget
Spot cost minus forward cost−PLN 20,000 or +PLN 20,000

At spot 4.1000 the forward is PLN 20,000 more expensive, yet still meets the budget. At 4.5000 it avoids a PLN 20,000 higher unhedged cost and a budget overrun. The comparison changes sign while the predefined budget objective remains met. That difference is not an extra cash payment added to the deliverable forward: the company pays PLN 430,000 and receives EUR 100,000. This does not make every hedge suitable; an uncertain invoice, wrong amount or changed payment date would require a different assessment, while the forward still creates its own obligation.