Until the mid-1990s, Spain was seen as the paradigm of high unemployment among developed countries. But things have changed considerably. From 1995 to 2006, the unemployment rate has fallen by 15 percentage points, from 22% to 8%, while inflation remained subdued around 3.5%. Thus, as depicted in Figure 1, there have been remarkable changes in the Spanish Phillips curve, which has shifted dramatically inwards while becoming almost flat. It is well known that these trends have been shared by other economies, notably the US, in the context of what Chairman Bernanke has termed the “Great Moderation”. However, the fall in Spanish unemployment has been much more pronounced than elsewhere – 12.5 points larger than the average 2.5 percentage-point drop in the Euro area – while inflation has remained stable being one percentage point higher than the Euro-area average.
Most of the standard stories proposed to explain the shifts of the Phillips curve in other countries do not fare well when analysing the Spanish case. First, while structural unemployment (the NAIRU) has clearly fallen, it is difficult to identify key labour or product market reforms, relative to those implemented in other economies, which could explain such a favourable evolution. Next, one of the most popular explanations for the improved inflation-unemployment tradeoff in the US – namely a rise in productivity growth when workers have backward-looking real-wage aspirations (Ball and Moffitt, 2002) – does not fit the Spanish experience either. Indeed, over the last decade, productivity growth has fallen in Spain, being among the lowest in the EU. Likewise, arguments based on the effects of increasing trade openness are unable to explain why equally large shifts in the Phillips curve have not taken place in other EU countries subject to the same trade integration and global competition patterns. Finally, a common monetary policy probably has contributed to the flattening of the Phillips curve, as low inflation expectations became better anchored. Still, why inflation did not surge vis-à-vis the Euro area in the face of such a large reduction in unemployment remains puzzling.
Somewhat surprisingly, none of the available studies dealing with the evolution of the Spanish Phillips curve have addressed the role played by the most fundamental change affecting the Spanish labour market during this period, namely the immigration boom that has taken place since the mid-1990s. As explained in a recent Vox column by one of us, while the proportions of foreigners in the Spanish population and labour force hardly reached 1% in 1995, they have since soared to about 10% and 14%, respectively.1 Thus, it seems worthwhile exploring whether this phenomenon can provide the missing piece in the puzzle.
In a recent CEPR Discussion Paper, we analyse the consequences of immigration for the joint behaviour of unemployment and inflation.2 So far, this topic has drawn little attention in the literature on the Phillips curve, the exceptions being the theoretical contributions by Razin and Binyamini (2007) and Engler (2007), where it is shown that higher migration flows can induce a flatter Phillips curve by changing both the aggregate labour supply and labour demand elasticities. Our approach stresses several labour-market channels through which immigration can affect inflation. In effect, to the extent that wages are differently determined for natives and immigrants –for instance, if immigrants are less well represented by unions than natives– or insofar as the marginal rate of substitution between consumption and leisure is different for each group – immigrants tend to be more mobile and more willing to take low-paid jobs than natives – expected real marginal costs of producing output can fall as immigration increases, leading to lower inflation for a given unemployment rate than in the absence of foreign workers.
In the seminal work by Galí and Gertler (1997) on the foundations of the so-called New Keynesian Phillips curve, when firms can only reset prices infrequently, the present discounted value of future real marginal costs faced by price-setting firms is the key determinant of inflation. In a competitive labour market, this effect is captured by the expected evolution of the labour share of GDP. However, this does not describe the Spanish labour market, where real wage rigidities are present in the wage-setting process through indexation. By considering a New Keynesian Phillips curve where real wages are sluggish, as in Blanchard and Galí (2007), we find that the conventional specification of the expectations-augmented Phillips curve (where current inflation depends on future and lagged inflation plus the unemployment rate and imported inflation) can be recovered. By distinguishing between native and immigrant workers –which are assumed to be imperfect substitutes with different preferences towards consumption and leisure, and different bargaining power (lower for immigrants) in non-competitive labour markets – two new variables play a role in explaining the short-run inflation-unemployment trade-off: the gap between the unemployment rates of the two types of workers and the immigration rate in employment.
The intuition behind these effects is as follows. First, insofar as the unemployment rate of immigrants is higher than that of natives (a realistic assumption supported by the evidence, see Figure 2) and as long as the former have a more inelastic labour supply than the latter (immigrants have higher participation rates), immigration reduces the expected real marginal cost faced by firms. Think of a negative labour demand shift: wages will fall more the less elastic labour supply is, leading to lower inflationary pressure. Further, given a lower bargaining power than natives, a higher unemployment rate among immigrants leads to lower wage markups. Of course, the opposite would happen if, in the future, the unemployment rate of immigrants fell below that of natives, as long as their labour supply remained less elastic. Finally, it can be shown that the size of these two effects hinges on the immigration rate in the economy: the higher the immigration rate, the larger the inflation reduction brought about by the unemployment rate gap. In other words, these effects on the real marginal cost will be larger in economies with rapidly increasing immigration rates, as in Spain.
Taking this new specification of the New Keynesian Phillips curve to the Spanish data, we find both an excellent fit and sensible results in terms of the size and the sign of the effects associated to the determinants of inflation –much better than with the standard specification. We calibrate alternative counterfactual scenarios to examine what would have happened if the immigration boom had not existed. Our results are striking: (i) the slope of the Phillips curve would have been almost four times steeper; (ii) the intercept would have been higher by about half a percentage point per year, and (iii) the average annual inflation rate would have been 2.2 percentage points higher. To these labour market effects, one could add the moderating effects of immigrants on inflation through their direct effect on product markets, as recently shown by Lach (2007), who stresses their role as “new consumers” with higher price elasticities and lower search costs than natives.
Overall, these results reveal that the inward shift and the flattening of the Phillips curve due to immigration imply that demand shocks and policy mistakes do not show up in large movements of inflation as long as immigrants take time to integrate. Although it is too early to detect such evolution in the case of Spain, if immigrants’ labour supply comes closer to that of natives and inflation remains above target, a deeper slowdown or increasing immigration flows will be needed to bring it down.
Ball, L. and R. Moffitt (2002), "Productivity Growth and the Phillips Curve", in A. Kruger and R. Solow (eds.) The Roaring Nineties: Can Full Employment Be Sustained?, Russell Sage Foundation.
Blanchard, O. and J. Galí (2007), "Real Wage Rigidities and the New Keynesian Model", Journal of Money, Credit, and Banking (forthcoming).
Engler, P. (2007), "Gains from Migration in a New-Keynesian Framework", mimeo, Free University of Berlin.
Galí, J. and M. Gertler (1999), "Inflation Dynamics: A Structural Econometric Analysis", Journal of Monetary Economics 44, 195-222.
Lach, S. (2007), “Immigration and Prices,” Journal of Political Economy 115, 548-587.
Razin, A. and Binyamini (2007), "Flattening of the Short-run Trade-off between Inflation and Domestic Activity: The Analytics of the Effects of Globalization", mimeo, Tel Aviv University.
2 Bentolila, S., Dolado, J. and JF Jimeno (2007) “Does immigration affect the Phillips Curve?. Some Evidence from Spain” CEPR. DP. 6604