Aggregating performance returns for a portfolio

Views:

Related resources:

You can ask LUSID to aggregate performance returns for a portfolio over a particular period, for example the preceding 5 days, the year to date, since fund inception, or all of these.

To aggregate performance for a transaction portfolio, you must first upload your own simple performance returns to record the net gain or loss of the portfolio, either per day (the default) or per month.
To aggregate performance for a reference portfolio, you do not need to upload your own performance return data but you must ensure quotes for each constituent instrument in the portfolio are uploaded to the LUSID quote store.

Uploading a simple performance return for a transaction portfolio

Note: You do not need to upload simple performance returns for a reference portfolio.

To create or update a simple performance return for a transaction portfolio, obtain an API token and call the UpsertPortfolioReturns API, specifying the scope and code of the transaction portfolio and then the returnScope and returnCode of a particular set of simple performance returns, to which you can add as many individual returns as you like. You should supply a PerformanceReturn object in the body of your API request with the following data:

Field	Status	Explanation
`effectiveAt`	Required	An effective datetime marking the end of the return period.
`rateOfReturn`	Required	The net gain or loss of the investment over `period`.
`openingMarketValue`	Optional	The market value at the start of `period`. Though optional, you should specify either this or `closingMarketValue`; do not omit both. LUSID uses the formula below this table to calculate missing market values.
`closingMarketValue`	Optional	Specify the market value at the end of `period`. Though optional, you should specify either this or `openingMarketValue`; do not omit both.
`period`	Optional	The default is `Daily`; this constitutes a return for one day. Specify `Monthly` to specify a return for one month.

You can specify any number you like for the rateOfReturn. For example, you could calculate it as a simple percentage difference between the opening and closing market values:

rateOfReturn = (closingMarketValue – openingMarketValue) / openingMarketValue

Note an individual simple performance return may not have either an openMarketValue or a closingMarketValue (these fields are optional). If no openingMarketValue is provided, then LUSID calculates it as follows:

If it is the first performance return, then:
- If rateOfReturn is 0, then openingMarketValue is 0 as well.
- If rateOfReturn has a value then openingMarketValue is calculated using the formula:
```
openingMarketValue = closingMarketValue / (rateOfReturn + 1)
```
If it is not the first performance return, then openingMarketValue is the previous closingMarketValue.

If no closingMarketValue is provided, then LUSID calculates it using the formula:

closingMarketValue = (rateOfReturn + 1) * openingMarketValue

Note you can:

List all the simple performance returns you have uploaded to a transaction portfolio using the GetPortfolioReturns API.
Delete a particular simple performance using the DeletePortfolioReturns API.

Calculating aggregated performance return for a transaction or reference portfolio

Call the GetPortfolioAggregatedReturns API, passing in the scope and code of the portfolio and, optionally, fromEffectiveAt and toEffectiveAt dates to restrict the returns to aggregate.

Note: If the portfolio is a composite portfolio, LUSID automatically aggregates performance for all constituent portfolios. Note if you want the dispersion metric you should call the GetAggregatedReturnsDispersionMetrics API instead, providing the composite has at least six constituent portfolios and each has been in for the whole year. If you want to analyse how the composite return has been calculated afterward, call the GetCompositeBreakdown API.

You should supply an AggregatedReturnsRequest object in the body of your API request with the following data to control how LUSID aggregates returns:

Field	Status	Explanation
`metrics`	Required	You must provide at least one `PerformanceReturnsMetric` object; see the following table.
`returnIds`	Required for a transaction portfolio	This is the `returnScope` and `returnCode` of the set of simple performance returns to aggregate.
`recipeId`	Optional	This is the `scope` and `code` of a recipe from which to get FX rates to convert returns into the base currency of the portfolio (including inferring from an inverse FX rate if the appropriate parameter is set). If not specified, this is the built-in default recipe.
`compositeMethod`	Optional	The default is `Asset`; if the portfolio is a composite, a weighted average of all constituent portfolios is used to determine the result. Specify `Equal` to determine that all the constituent portfolios in the composite are weighted equally.
`period`	Optional	The default is `Daily`. You can change this to `Monthly`. Note if you specify `Monthly` but have no monthly performance returns then daily returns are used to calculate a monthly return if sufficient of the former are available.
`outputFrequency`	Optional	The default is `Daily`; you can change this to either `Weekly`, `Monthly`, `Quarterly`, `HalfYearly` or `Yearly`. See also the `PERIODRETURN` constant window in the following table.
`alternativeInceptionDate`	Optional	The date from which to include performance returns if this is different from the date when returns begin. Can be a date string or a portfolio property. Defaults to the actual inception date.
`holidayCalendars`	Optional	The holiday calendar code(s) that should be used to determine the date schedule as a comma-separated list (for example `CoppClark`).
`currency`	Optional	Converts the results to the specified currency. If omitted, results are in the portfolio currency. Can be a ISO 4217 currency code string or a portfolio property.
`runMode`	Optional	Defaults to `ReturnData`, in which case LUSID calculates results for the returns uploaded to the system. Specify: `AllDays` to calculate results for all dates including weekends (if there is no return for a particular date, it is treated as 0 rather than omitted). `WeekDays` to calculate results for all dates excluding weekends (if there is no return for a particular date, it is treated as 0 rather than omitted). If `period` is `Monthly`, `MonthEnd` to use the last return found for a particular month as the return for the last day of the month, rather than not returning a result.

For the metrics field (above), supply at least one PerformanceReturnsMetric object containing the following data:

Field Status Explanation

type

Optional

Defaults to Return. Specify:

Volatility to cause LUSID to calculate based on the statistical dispersion of returns (how much an asset's price swings around the mean price).
IndicativeAmount to calculate based on investing the seedAmount (or wealth index; see below).

window

Required

Specify the window to aggregate. This should be a string constituting either a constant window, a multi-X metric, or a Since metric.

Constant window

String	Explanation
`PERIODRETURN`	Aggregation depends upon the value of `outputFrequency` (see the table above): If `outputFrequency` is `Daily`, this is a ‘1 day return’. If `outputFrequency` is `Weekly`, this is a ‘1 week return’. If `outputFrequency` is `Monthly`, this is a ‘1 month return’. If `outputFrequency` is `Quarterly`, this is a ‘quarter return’. If `outputFrequency` is `HalfYearly`, this is a ‘half year return’. If `outputFrequency` is `Yearly`, this is a ‘1 year return’.
`WTD`	Aggregates returns for the week to date, where Monday is the first day of the week.
`MTD`	Aggregates returns for the month to date.
`QTD`	Aggregates returns for the quarter to date.
`YTD`	Aggregates returns for the year to date.
`INC`	Aggregates returns since inception.

Multi-X metric

Since months may be of varying length, there is special handling for the last day of the month. Month-end dates look back to prior month-end dates. So, one month rolling for the 28th February will look back to the 31st of January, not the 28th January.

String	Explanation
`XD`	Aggregates returns for X days (eg. `3D`)
`XW`	Aggregates returns for X weeks (eg. `3W`)
`XM`	Aggregates returns for X months (eg. `3M`)
`XY`	Aggregates returns for X years (eg. `3Y`)

Since metric

String	Explanation
`Since(yyyy-mm-dd)`	Aggregates returns since a particular date (eg. `Since(2012-12-20)`)
`Since(YYYY-mm-dd)`	Aggregates returns since the most recent occurrence of a day in a year (eg. `Since(YYYY-12-20)`)
`Since(PropertyKey)`	Aggregates returns since the value of a particular property (eg. `Since(Portfolio/Test/ReturnDate)`)

allowPartial Optional Defaults to False; returns are only aggregated if the date range is fully covered by the return history. For example, a one year return is only aggregated if the portfolio has returns posted for at least one year.

Specify True to aggregate returns from the current state of the return history. Note results will be partial.

annualised Optional Defaults to False. Specify True to annualise the result.

withFee Optional Defaults to False. Specify True if period is Monthly to include model fees in the performance calculation, and apply the monthly rate of fees using the Portfolio/default/ModelFees system property. Please note this is only supported for monthly returns.

seedAmount Optional Specify a decimal number. The result will be calculated using the formula seedAmount * (1 + window-value), providing type is set to indicativeAmount.

alias Optional Specify a name to display for the metric in the response.

Keywords: portfolio, return, aggregated return, aggregation, performance