Treasury staff insight

FEU Special Topic: Decomposing inflation into supply and demand drivers

The Treasury

Staff and teams are writing in their individual capacity and the views are not necessarily a "Treasury" view. Please read our disclaimer.

Special Topic: Decomposing inflation into supply and demand drivers

The rise in inflation in New Zealand reflects both supply and demand factors. On the supply side, the pandemic disrupted labour supply, creating bottlenecks in production, which the war in Ukraine exacerbated, while poor domestic weather has also played a role. As economies adapted to the pandemic and reopened, demand recovered, boosted by stimulatory macroeconomic policy, and the sources of inflation broadened to include more services.

The key drivers of New Zealand’s inflation have been the food, housing and transport groups, and the persistence of high inflation has caused an increase in inflation expectations and seen inflation spread to other components (Figure 1).

Figure 1: Composition of CPI inflation

Feu 31mar 23 special topic fig 7

Source: Stats NZ

Estimating supply and demand drivers

Economic models can illustrate the contributions to inflation from supply and demand factors. To analyse the contribution of supply and demand factors, one approach is to look at the behaviour of inflation and sales volumes by spending category, and aggregate these to form a view about overall inflation. Based on the standard assumption of an upward sloping supply curve and downward sloping demand curve for each spending category, inflation can be decomposed into supply or demand drivers using time series models[1].

This approach suggests that supply and demand drivers that the model can explain have contributed roughly equal amounts to annual inflation in the past year, although different models give different estimations (Figure 2). About one-third of inflation is unable to be attributed to either supply or demand. The surge in inflation from 2021 was initially demand driven by stimulatory monetary and fiscal policy, and rapidly increasing house prices which fuelled economic activity. However, from the second half of 2021, supply-side factors drove the continued accelerated in inflation. Contributions from supply and demand drivers have been relatively stable across 2022.

Figure 2: Decomposition of CPI inflation*

Feu 31mar 23 special topic fig 8

Source: Stats NZ, the Treasury

*The ambiguous bar represents the sum of components where the model errors are too close to zero to ascertain whether the shock is supply or demand driven. This accounts for the 20% smallest price and volume residuals.

Stats NZ suppress the education and health groups for quarterly household consumption expenditure which mean that not all of CPI inflation can be explained by the model. This is reflected in the bars not adding to CPI.

Data judgements affect the model output

The components of the CPI and household consumption expenditure (HCE) data have key differences which require specific judgments in mapping price and volume series together. Since 2000, the average difference in annual inflation between the CPI and HCE deflator is about 0.5 percentage points. The Reserve Bank of New Zealand’s policy target uses the CPI, and while judgements are required to map HCE groups to the CPI groups, this process provides valuable information for policymakers.

The judgments required to map the HCE components to the CPI components may affect the model output. When using the HCE price and volume series, the contribution from the supply factors to annual growth in the HCE deflator are larger, accounting for about 60% of explained inflation (Figure 3). These results are more consistent with international comparisons and highlight potential sensitivities of the model to the judgements made.

Figure 3: HCE deflator decomposition

Feu 31mar 23 special topic fig 9

Source: Stats NZ, the Treasury

Housing and food have made outsized contributions to CPI inflation

The housing and utilities group has been the largest driver of CPI inflation in recent quarters. Large differences in the definition of the housing group of the CPI and HCE mean significant judgements are required to map these price and volume components to one another.

The driver of housing and utilities group inflation in the CPI is typically hard to determine, although there are episodes throughout the last decade that highlight demand and supply side influences (Figure 4). The supply of new housing generally has long response times which may be reflected in the model’s limited ability to explain the majority of the group’s inflation. The approach does, however, show evidence of supply-side drivers from the middle of 2021, consistent with the timing of widespread materials and labour shortages. Supply-side drivers have since eased, bringing down the rate of annual inflation, although it remains elevated.

Figure 4: Home ownership subgroup

Feu 31mar 23 special topic fig 10

Source: Stats NZ, the Treasury

The food group has also made a large contribution to inflation in recent quarters, and supply-side factors have been the key driver (Figure 5). Poor growing conditions, labour shortages and rising input costs have constrained output. The impacts of Auckland flooding and Cyclone Gabrielle are expected to further constrain the supply of fresh foods, keeping food inflation elevated over the first half of 2023.

Figure 5: Food group

Feu 31mar 23 special topic fig 11

Source: Stats NZ, the Treasury

Inflation has spread to other components

Services inflation continues to increase in New Zealand likely the result of two factors (Figure 6). As COVID-19-related restrictions have eased, which heavily constrained the services sector, consumers have switched consumption back from goods, generating strong demand pressure. Additionally, high inflation over the pandemic and tight labour markets have flowed into higher inflation expectations and strong wage demands.

Figure 6: Services

Feu 31mar 23 special topic fig 12

Source: Stats NZ, the Treasury

Cross country comparisons

The results for New Zealand show lower supply-side contributions to inflation than estimates for the US and Australia. In the US, supply-side drivers account for about 60% of the annual change of the PCE deflator that the model can explain (Figure 7).[2] Similarly in Australia, the Reserve Bank of Australia estimated that about two thirds of explained annual inflation in 2022 was driven by supply factors.[3]

The similar experience of New Zealand likely reflects the global nature of some of the supply-side shocks that have occurred across the pandemic. However, the lower impact of supply-side shocks may reflect that New Zealand has been more insulated from some energy inflation from the war in Ukraine than other countries. The larger contribution from demand factors may also reflect the relatively large policy response in the early stages of the pandemic.

Figure 7: US headline PCE deflator

Feu 31mar 23 special topic fig 13

The outlook for inflation

The results outlined in this note highlight the confluence of factors that have contributed to high inflation and help policymakers understand how inflation might evolve over the coming quarters.

At HYEFU, the Treasury forecast CPI inflation had peaked in late 2022 and that it will gradually decline over the coming years. The key factors behind this are the resolution of global supply-side disruptions from across the pandemic that are resulting in lower freight costs and reduced pressures on goods and commodity prices. The impact of higher interest rates across the world is also lowering demand-side pressures on capacity and prices.

There are signs that higher interest rates domestically are contributing to a slowdown in demand and activity, particularly in interest-rate sensitive sectors, although businesses continue to report intense cost pressures, which they are passing on. There remains considerable uncertainty about the domestic supply outlook, reflecting constraints from recent weather events. The Treasury’s initial estimate suggests these events will contribute 0.4 percentage points to CPI inflation over the first half of 2023, although these effects are expected be temporary.

  1. [1] Using Shapiro’s method, we estimate two equation VAR models for each component using the logs of price and volume measures with four lags to recover the reduced form residuals. If price and volume residuals have the same sign, this suggests a demand shock. If the price and volume residuals have opposite signs, this suggests a supply shock.
  2. [2] Shapiro, Adam Hale. 2022. "Decomposing Supply and Demand Driven Inflation," Federal Reserve Bank of San Francisco Working Paper 2022-18.
  3. [3] See the Reserve Bank of Australia’s February 2023 Statement on Monetary Policy.
Join the discussion on LinkedIn