Fiscal risk analysis at the OBR

Fiscal risk analysis at the OBR
Robert Chote
6th Annual meeting of OECD PBOs and IFIs
Jerusalem, 1 April 2014
Context: role of the OBR
• “examine and report on the sustainability
of the public finances”
• Four main tasks:
Produce five-year forecasts for economy and public finances
Judge progress towards government’s targets
Scrutinise costing of tax and welfare policy measures
Assess long-term sustainability of public finances
Levels of fiscal risk
• Risk of deviation from forecast
• Risk of missing stated targets
• Risk of long-term unsustainability
– And eventual need for some policy change
• Risk of crisis
– And sudden need for big policy change
Medium term forecast/targets
• Required to assume unchanged policy
– Ignore political risk but not delivery risk: fuzzy line
• And to judge if government has a ‘better
than 50%’ chance of meeting targets
• Implies pass/fail test on median forecast
• But important to recognise uncertainty
Sources of fiscal forecast deviation
• Economic determinant forecast errors
• Fiscal forecasting (‘technical’) errors
– E.g. yield from tax disclosure agreements
– E.g. forestalling of pre-announced tax rate changes
– E.g. local authorities’ spending behaviour
• Policy changes
• Classification changes
– E.g. change to net debt definition
Explaining forecast errors ex post
Illustrating forecast/target uncertainty
• Past forecast errors
• Sensitivity to key economic determinants
• Scenario analysis
Probabilities based on past errors
Cyclically-adjusted current budget balance
Past errors imply c.75% chance of achieving target of
structural current budget balance five years ahead
Mechanical sensitivity analysis
• Bigger/smaller output gap
• Faster/slower GDP growth
• Higher/lower government borrowing costs
• Higher/lower cyclical adjustment coefficients
Selected scenario analysis
• Used to highlight key debates/critiques
• March 2014: what if interest rates rise faster?
– Response to faster recovery (‘good’)
– Response to rise in risk premia (‘bad’)
• Past examples
– Spike in oil prices
– Euro-zone crisis (took OECD scenario)
More focus on economic risks?
• Easier to summarise in sensitivity and scenario
analysis – and in forecast narrative
• In March 2014 we highlighted as risks:
Volatility from tighter global monetary policy
Danger of renewed euro-zone instability
Escalation of tension in Russia/Ukraine
Productivity growth failing to recover
Growing household borrowing and indebtedness
• Fiscal forecasting risks more specific
– Some tax bases hard to forecast/proxy
• e.g. income and house price distribution
– Behaviourial/administrative uncertainty
• e.g. trends in miles driven or disability assessments
– Also look at contingent liabilities from WGA
Quantifiable contingent liabilities
• WGA only available with a lag, so also note new liabilities
• In June 2013 highlighted greater use of guarantees
– Probably ‘remote contingent liabilities’ but also correlated if economy tanks
Unquantifiable contingent liabilities
Long-term uncertainty/risks
• Central projection based on ‘unchanged’ policy
• Calculate debt trajectory and fiscal gaps
• Sensitivity analysis
Fiscal position at end of medium-term forecast
Long-term relationship between interest rate and growth rate
Demography: ageing / net migration flows
Health spending
• Also issues in tax revenue sustainability
North Sea oil receipts
Motoring taxes
Globalisation and corporation taxes
Retirement behaviour
Risks in policy costings
• Key judgements in policy costings set out with
each Budget / Autumn Statement
• OBR annex highlights key uncertainties, usually
around behaviour or identifying tax base
• Anti-avoidance measures often popular and
particularly uncertain in their effect
– Expected attrition taken into account explicitly
• Sometime policy costings not certified/included
because details, timing or approval not certain
– E.g. assets sales where time/price not known
– Required to identify these as risks