Spatial inequality is central when analysing economic performance across regions and countries. Governments worry about wealth and prosperity concentrated in a few ‘lucky’ areas. In the UK, the dominant position of London and the absence of a sizeable contender fuels the debate on geographical rebalancing. In Denmark, the government is currently expanding a relocation policy (started two years ago) with the aim of reducing spatial inequality between Copenhagen and other parts of the country.

One possible way of obtaining geographical rebalancing is through relocation policies. Since the 1950s, governments have frequently used relocation programs to address unemployment in declining regions. Relocation advocates claim that the policies trigger local multiplier effects: the arrival of public sector jobs in an area increases demand for locally produced goods and services. Opponents stress that newly created jobs merely crowd out existing ones: upward pressure on local rents increases production costs with negative consequences for businesses. With limited evidence collected so far, the jury is still out. Our work tries to disentangle the puzzle – although more evidence is needed before we can confidently rule who is right.

In a recent SERC discussion paper, we study the impact of the German government move from Bonn to Berlin in the 1990s in order to understand the interaction between public and private employment within a local labour market. Our goal is to identify which effect (multiplier or crowding-out) prevailed in the German government case and then to quantify the size of the effect.

We do find evidence of a local multiplier effect: the arrival of 100 public sector jobs into one of the 190 Berlin postcode areas generates about 55 additional jobs in the private sector. There is also evidence of spillovers: the inflow of 100 public sector jobs triggers job creation (36 new jobs) in areas neighbouring the relocation site. These effects come through service sector jobs, while manufacturing employment is not influenced at all. Additionally, we find that agglomeration effects are highly localised within the first few kilometres of a relocation site, disappearing sharply over distance.

Where do the service jobs come from? When we split the service sector by sub-group, the most important channel for the employment multiplier is business-to-business activities: out of a total of 56 jobs, 27 jobs are created in the business sector, i.e. about 48 per cent of all newly created jobs. This is followed by trade (4.4 jobs), real estate (3.2 jobs) and personal services (1.6 jobs). Weaker estimates are then reported for media services (7.9 jobs), transport activities (7.6 jobs) and tourism (2 jobs).

The policy appears to be a success. It boosts job creation in the private sector, especially in services. Does this mean we should encourage other nations to implement similar relocation policies? Not so sure about that. Obviously, there is something to learn from the German case. But still, we need to be aware of the specifics of this historical episode.

This was a massive relocation: about 15,000 government-related positions were relocated from Bonn to Berlin within a relative short period of time (1999-2001). Additional 10,000 positions related to diplomacy, the media, political parties and interest groups followed suit. At the same time, Berlin experienced significant outflows of public sector jobs linked to the redistribution of agencies across the federal states that used to be part of East Germany. Adding up positive and negative moves resulted in a net gain of about 18,000 jobs for Berlin.

This massive relocation targeted high-profile jobs involving top-level civil servants. Following the Berlin-Bonn Act (1994), a ‘fair division of labour’ between Berlin and Bonn was decided together with compensatory measures for the former capital. Six ministries kept their first seat in Bonn and got a second seat in Berlin; nine ministers took their first seat in Berlin and kept their second seat in Bonn. What is unique about this relocation is that high-profile positions were moved out of a former capital into a new capital, with no associated loss of prestige or career prospects. Although we don’t deny the individual and family costs of such moves, they didn’t have detrimental effects in terms of career and job prospects. Again, we don’t know how many people moved with their job but we expect the number to be high.

Such as massive relocation targeted a few areas: 38 out of 190 Berlin postcodes received the relocated jobs; 30 were affected by a positive employment shock; 8 by a negative shock. Relocated employment per postcode ranges from -813 employees in the postcode that lost an important federal institution to about 5,200 employees in a postcode in ‘Berlin Mitte’, the most centrally located district in Berlin, where several ministries and embassies clustered. Moreover, location decisions were largely determined by historical factors, i.e. by the desire of the government to occupy (where possible) buildings of historical importance.

Still, our analysis has limitations: it is a partial analysis. It identifies the average policy impact within a given postcode; it doesn’t allow us to compare Berlin affected by the program with Berlin under a non-relocation scenario. Our analysis also fails to capture a further important aspect of the relocation program: the effects of relocated government employee residential choices. Thus, we cannot study localised effects on the housing market or changes in consumption patterns.

This historical episode shows an example of successful relocation: however one swallow does not a summer make. Naturally, move evidence needs to be gathered. Economists have been working on the subject both in the UK (Faggio and Overman, 2014) and abroad (Jofre-Monseny, et al., 2016), but defining the parameters continues to be elusive. Undoubtedly, understanding these issues will provide possible answers to contemporary relocation challenges: the announced move of 500 Channel4 jobs out of London; the post-Brexit move of UK-based EU agencies; and the ongoing debate about the spatial dimension of immigrant resettlement programs. We will have to wait a little longer before we out the truth.



Giulia Faggio has a BSc and an MSc in Economics from Bocconi University, Italy; and a PhD in Economics from K.U.Leuven, Belgium. She also attended the Advanced Studies Program in International Economic Policy Research (ASP) at the Kiel Institute for World Economics, Germany. She is a lecturer in economics at City, University of London and a research affiliate at LSE’s Spatial Economic Research Centre. On this subject, she has been looking at the impact of other UK relocation programs (i.e. the Lyons Review).

Teresa Schlüter is an economic analyst at the German advisory council on global change where she works on sustainable urban development. She holds a PhD from the London School of Economics in Economic Geography. Her interests lie in the areas of urban and labour economics, inequality, innovation policy and micro-econometrics.


Philipp vom Berge studied economics in Regensburg and graduated in 2007 (degree “Diplom-Volkswirt”). He was a research assistant at the Department of Economics of Regensburg University until 2012. He joined the Research Data Centre of the Federal Employment Agency as a researcher in 2011. His research interests lie in regional and labour economics.