Publications

"Average Inflation Targeting and the Interest Rate Lower Bound", with Taisuke Nakata and Sebastian Schmidt, European Economic Review, Vol. 152, February 2023, [WP version, Appendix, CEPR WP, ECB WP,  BIS WP]

Under conventional inflation targeting (IT), the lower bound on nominal interest rates gives rise to a systematic downward bias in inflation that substantially reduces welfare. Using two variants of a New Keynesian model, we investigate whether a monetary policy strategy that aims to stabilize an average inflation rate—rather than a period-by-period inflation rate—leads to better outcomes. With rational expectations, price level targeting (PLT)—the limiting case of average inflation targeting (AIT)—is optimal, yet AIT with sufficient history dependence reaps most of the benefits of PLT. With boundedly rational expectations, PLT is no longer optimal unless the degree of bounded rationality is small. When deviations from rational expectations are sufficiently large, outcomes can be worse under PLT or AIT with strong history dependence than under IT. Finally, for both variants of the model, inflation conservatism improves welfare by eliminating the deflationary bias without invoking history dependence. 

Working Papers

"Shrinkflation" [WP version, SSRN]

This paper studies the macroeconomic relevance of product quantity adjustment, i.e., changes in product size, weight or quantity. I establish several facts from U.K. CPI micro price data from 2012-2023: (i) Quantity changes are relevant for around 35% of the consumption basket. (ii) Each month, up to 0.7% of goods in the CPI experience quantity  changes, with considerable variation across sectors and products (e.g. up to 3% in the Food sector or up to 60% for products like chocolate). (ii) The majority are quantity reductions, particularly those resulting in “downgrades,” where the quantity decreases relative to the price. (iii) The frequencies of quantity reductions and downgrades are strongly procyclical, moving positively with CPI and PPI inflation and negatively with the unemployment rate. Conversely, quantity increases and upgrades are mostly acyclical. (iv) The frequencies of price and product quantity adjustment are largely unrelated. These facts highlight the systematic use of “shrinkflation” – whereby unit prices increase due to quantity reductions – as a margin of adjustment in response to aggregate fluctuations.

"Temporary Sales and Cyclicality" [WP version, SSRN]

How do temporary markdowns (”sales”) impact aggregate price flexibility? This paper provides novel evidence from U.K. CPI price micro data that sales matter for regular price adjustment over the business cycle. Around 50% of sales occur right after or before a regular price change, as if they are strategically placed to conceal the price hike or highlight the price cut. These ”strategic sales” are strongly countercyclical, peaking during the Great Recession and the Covid-19 pandemic. This suggests that sales are used as a means to facilitate cyclical adjustments of regular prices. Strategic sales are linked to up to 15% of regular price changes and are the main driver of aggregate sales dynamics.

"Inflation Targets and the Zero Lower Bound" [WP version, SSRN] 

Does a higher inflation target help to reduce the risk of hitting the zero lower bound (ZLB) on nominal interest rates? Recently, higher inflation targets for central banks have been proposed to allow for more "room-to-manoeuvre" in deep recessions. Advocates of this proposal suggest an inflation target of 4% to reduce the risk of hitting the ZLB. I show that the opposite may happen: a 4% inflation target can, in fact, increase the risk of hitting the ZLB compared to a 2% inflation target. Using a standard New Keynesian model, a higher inflation target changes the price-setting behavior of firms in a substantial way. Specifically, firms become more forward-looking, inflation is more volatile and, thereby, the nominal interest rate fluctuates more. I show that even with more "room-to-manoeuvre" for the nominal interest rate due to a higher inflation target, the higher volatility of the nominal interest rate implies that the economy ends up – on net – more often at the ZLB. 

"Are Consumption Tax Cuts Expansionary in a Liquidity Trap?" [WP version]

Do temporary value-added tax (VAT) cuts stimulate aggregate consumption? I show that the canonical New Keynesian model predicts that they are expansionary in normal times but contractionary in deep recessions (i.e. when an effective lower bound on nominal interest rates is binding). A potential issue is that standard models only account for consumption of non-durable goods. However, in countries that levy VAT, consumer durables typically represent roughly 40% of total consumption expenditures on goods and services that are subject to the VAT. I allow for consumption of durable goods in the New Keynesian model and now find that the previous results are completely overturned: temporary consumption tax cuts have large positive macroeconomic effects both in normal times and in a liquidity trap. The reason is that purchases of durable goods are highly intertemporally substitutable - consumers will stock up on storable goods when prices are currently low. But most interestingly, I observe that the boom in the durable goods sector spills over to the non-durable goods sector. The VAT cut becomes expansionary for both non-durable and durable goods consumption. The findings of this paper suggest that it is important to distinguish between different types of consumption goods to study the aggregate effects of consumption tax changes. 

"Revisiting the Comovement Puzzle"

In standard two-sector monetary business cycle models, the Comovement Puzzle describes the observation that production and employment in different sectors move in opposite directions after a economy-wide shock. This is clearly at odds with empirical evidence. The reason for the Comovement Puzzle roots in two key assumptions in these models: Fully flexible durable good prices and free movement of resources across sectors. In this paper, I introduce wage rigidities in an otherwise standard two-sector New Keynesian model. Wage rigidities reduce the fluctuations in marginal cost of both non-durable and durable good producing firms. The latter generates a new source for price rigidities. The paper finds that a model specification with wage rigidities is able to replicate the comovement of production and employment in both sectors after a monetary contraction. From the estimated model using Bayesian techniques, it can be shown that some degree of price stickiness for durable goods and the presence of wage rigidity cannot be rejected.