Gilts Yield App

UK conventional gilt explorer and yield calculator with a small Python backend and a static frontend.

Browsable deployment: http://mazzo.li/gilts/

This project (including this file) is AI generated with human supervision.

What this project does

• Loads DMO-style "Gilts in Issue" spreadsheet data from ./gilts/*.xls.
• Extracts conventional gilts and exposes them over a JSON API.
• Computes:
  • Dirty price from clean price + accrued interest
  • Effective annual yield (XIRR-style)
  • Post tax return
  • Gross equivalent yield
• Serves a browser UI (when run with static serving enabled).

Repository layout

• gilt_yield.py: data parsing + yield math.
• gilts_webapp.py: HTTP API server (and optional static file serving for development).
• static/: frontend (index.html, app.js, styles.css).
• gilts/: input spreadsheet files.

Running the app

Start API + static UI:

python gilts_webapp.py --serve-static

Then open:

• http://127.0.0.1:5001/gilts

API-only mode:

python gilts_webapp.py

API

GET /gilts/api/gilts

Returns:

• today: server date (YYYY-MM-DD)
• active_rows: gilts with redemption date >= today
• past_rows: gilts with redemption date < today

POST /gilts/api/yield

Request JSON:

• isin: string, e.g. GB00...
• price: clean price per 100 nominal
• tax_rate: decimal fraction, e.g. 0.40
• purchase_date (optional): YYYY-MM-DD

Response JSON includes:

• accrued_interest_per_100
• dirty_price_per_100
• annualized_yield
• post_tax_return
• gross_equivalent_yield
• is_ex_dividend_period
• next_coupon_date

Yield model used

The main yield field is an effective annual IRR-like number:

• Uses dated cashflows and solves discount rate r such that NPV is zero.
• Time exponent uses days/365 (actual calendar day count divided by 365.0).
• Initial outflow is -dirty_price.
• Future inflows are coupon and redemption cashflows.

Taxed yield:

• Same solver and timing.
• Coupon component is reduced by tax_rate.
• Principal repayment is not taxed in the model.

Gross equivalent yield:

• post_tax_return / (1 - tax_rate)

Important assumptions and imprecisions

These are deliberate current-model choices or known limitations.

Convention differences from bond-market quoting

• This is not a GRY/DMO quote-convention calculator.
• It does not implement Act/Act ICMA + semiannual bond-equivalent quoting conventions.
• If you compare against quoted gilt yields from market data pages, small or sometimes meaningful differences are expected.

Day count and compounding

• Uses effective annual IRR with days/365.
• Leap years are not modeled with a changing denominator (e.g. 366 in leap years).

Coupon cashflow assumptions

• Coupon per period is modeled as coupon_rate / 2.
• Irregular first/long/short coupon stubs are not explicitly modeled.
• If a gilt has non-standard first coupon structure, yields may be off.

Ex-dividend logic

• Ex-div period start is computed as 7 business days before next coupon.
• "Business day" currently means weekday only (Mon-Fri), not UK bank-holiday-aware.
• Around holiday periods, ex-div boundary classification can be off by a day.

Tax modeling

• Single constant tax_rate is applied to coupon cashflows only.
• Capital gains are intentionally left untaxed to mirror UK gilt treatment; broader gains/losses, allowances, wrappers, and lot-level tax treatment are not modeled.
• This is a simplified comparison signal, not tax advice.

Data ingestion assumptions

• Input is expected to be DMO-like .xls format and headers.
• Parser is strict on required columns in the conventional gilts section.
• Header/schema changes in source spreadsheets can cause hard failures.

Scope assumptions

• Conventional gilts are parsed; index-linked section is not modeled for yield calc.
• Settlement defaults to workbook Data Date when no purchase date is provided.
• Merged dataset keeps the latest row per ISIN across local files.

Possible surprises in behavior

• Same isin + price can return different yields if purchase_date changes, because accrued interest and ex-div inclusion change.
• Near ex-div boundary dates, annualized_yield can jump due to coupon inclusion/exclusion.
• post_tax_return can diverge more than expected at high tax rates and short maturities.
• Missing or malformed rows in the source workbook are skipped or can trigger parsing errors depending on where they occur.

Security and operational notes

• API errors returned to clients are generic (Invalid request / Internal server error), while tracebacks are logged server-side.
• There is no authentication at the app layer.
• The built-in Python HTTP server is simple and intended for lightweight/self-hosted use, not high-throughput hostile traffic.

Not investment advice

This project is for exploration and comparison tooling. It is not financial, tax, or investment advice.