The main industrialized countries show clear signs of slowdown of economic activity and acceleration of inflation. In a word: stagflation. The first dramatic experience with stagflation on a global scale in the early 1970s changed profoundly common thinking about macroeconomic phenomena and policies. Today’s blueprints for policy makers have clear guidelines in the event of stagflation:
- the best monetary policy can do is to keep a low and stable inflation rate, given the structural conditions of the economy
- to this effect, spikes of excess inflation should be curbed by letting real interest rates grow (i.e. by raising the nominal policy rate above inflation)
- as a result, nominal wage setters should refrain from chasing current inflation being confident that it will quickly be curbed with small and transitory losses in real wages.
As matter of fact, the ongoing unprecedented escalation of oil and commodity prices is producing much less stagflation than in the 1970s. Nonetheless, at the first reappearance of the phenomenon, disorientation surfaces in the public opinion as well as among policy makers. The monetary authorities on the two sides of the Atlantic have clearly taken different stances. In the US, real interest rates on short maturities are negative with those on longer maturities being zero or barely positive. In the EMU real interest rates are positive at all maturities, and the ECB is sending clear signs that its stance will be kept in line with its mandate for price stability first. GDP outlook appears less severe in the US, and more severe in the EMU, than previously thought. Managing stagflation will still be more awkward and risky than indicated in the handbook for the modern central banker.
The basic problem with staglation is that it is a complex phenomenon involving a number of macroeconomic dimensions and interrelated agents. The simple recipe recalled above is best suited to nominal supply shocks. In practice: a sudden (unanticipated), generalized, and temporary soar of nominal production costs, with firms seeking to transfer higher costs to final prices. Yet this is emphatically not the story we are witnessing in the world economy. First, the nominal cost shock has to a large extent been unpredictable but it is by no means generalized. As shown by the ECB data in table 1, the shock has a distinct non-domestic dimension, entirely coming from oil and commodity prices which, for the most part, represent imported costs determined on world markets. Second, the ongoing upsurge of nominal imported costs will quite likely be of permanent rather than temporary nature – as testified by Mr. Trichet a few weeks ago. In fact, it is well documented that oil in the world is becoming more scarce; hence its real price (the real rent of pit owners) should rise.
These two facts suggest that industrialized countries are undergoing a sharp change in relative production costs, that is, a real supply shock. This gives rise to a situation which is totally different than that of nominal shocks, and which raises thorny problems in view of monetary policy implementation.
How far is the current oil price from its new equilibrium level, or in other words, how much additional imported inflation should oil users expect?
- How can the central bank keep the general price index on track if it has no control on its non-domestic component? Or for how long will the actual inflation rate of the general price index deviate from the benchmark?
- In theory, as a result of the new system of relative costs, real wages and profits in oil-user countries will hardly remain unchanged. Potential output, natural interest rate and NAIRU will not remain the same either. If central banks organize their policy with reference to these benchmarks, what is their new assessment?
It is both astonishing and worrysome that little (if any) echo of these fundamental issues can be found in the debate and official declarations about stagflation, especially in the EMU. To shed some light on these issues, let us consider the simplest short-to-medium-run determination of the output level. Profit-maximizing firms expand (reduce) output with respect to potential as long as nominal production costs grow less (more) than the general price index (the HICP, say). The HICP is under the responsibility of the central bank, which is commited to keeping it close to a given benchmark. Both nominal costs and the HICP consist of domestic and non-domestic items (the distinction refers to the markets where prices are set, not where goods are produced). Table 2 reports the composition of the HICP elaborated by the ECB. Typical non-domestic items are energy and non-processed food, which account for 17.4% of the overall index (though processed food prices, too, are largely affected by world commodity market conditions, as testified by their skyrocketing growth – see also ECB, Monthly Bullettin, June 2008). The ECB does not provide the decomposition of production costs, but as a first approximation we can say that it is roughly the same as that of the HICP.
Assuming that all nominal domestic costs follow nominal wage dynamics, the evolution of total nominal costs with respect to HICP inflation is given by
0.8xWI + 0.2xNDCI – HICP
HICP = 0.8xDPI + 0.2xNDPI
where WI = nominal wage inflation, NDCI = non-domestic cost inflation, DPI = domestic price inflation, NDPI = non-domestic price inflation.
Wages can only be set in periodic renegotiations. Thus the current (time t) nominal wage rate (and the related rate of increase WI) is the result of the latest renegotiation round (time t-1). At each round, wage setters seek to keep nominal wages in line with expected inflation. The latter can be viewed as a weighted average between the inflation benchmark indicated by the central bank (say 2%) and the observed inflation trend (the so-called “second-round effect”), i.e. WI at time t is given by:
- WI = wx2% + (1 – w)xHICP-1
The weight w captures various conditions affecting labour-market negotiations. On the one hand, it depends on the credibility of the inflation benchamark: w = 1 indicates that the 2% benchmark is fully credible (differences with HICP are regarded as small and transitory). On the other hand, w < 1 indicates that the 2% benchmark is not fully credible and that workers wish to protect their pruchasing power against inflation spikes that are neither small nor transitory. The more HICP dwells above 2%, the smaller will be w. The actual value of w, however, also depends on worker’s negotiation power with firms.
Is a neutral monetary policy possibile?
By neutral monetary policy I mean the ideal case in which the central bank keeps the HICP in line with its benchmark, while all underlying real variables are kept at their equilibrium level by market forces. This ideal state (ignoring trend growth) implies that real wages, real costs and output are constant (“vertical” aggregate supply). That is, it should be that 0.8xWI + 0.2xNDCI = 2%, and WI = 2% at the same time. Clearly, this is not possible unless NDCI is by fluke just 2%! To gauge the present order of magnitude of NDCI we may refer to the production price index in table 1, assuming that this index reflects cost dynamics. Thus we may posit that
- 0.8xWI + 0.2xNDCI = PPI = production price inflation
The figures for WI and PPI in table 1 indicate that the trend of NDCI over the last six months is about 16%.
Our simple algebra shows that, as said above, when NDCs are bound to rise in real terms (NDCI > HICP) a neutral monetary policy is no longer possible. This poses serious problems for the rational interplay between the central bank and wage setters, as we shall see in a moment. Various possible combinations of WI, NDCI and HICP give rise to different scenarios, the most interesting of which can be summarized in the following table
What does wage moderation mean, and does it pay?
Wage moderation is a central element in the post-1970s approach to stagflation in order to prevent an endless wage-price spiral. So far, this also seems the key concern in Frankfurt. But what does it really mean to, and implies for, workers?
The first and more popular meaning of moderation refers to the nominal wage increases, which should not exceed the medium-term benchmark set by the central bank. In our terms, w = 1. It is rational for workers to accept this pact with the central bank because it helps the central bank to keep inflation close to the benchmark, and as long as inflation remains close to the benchmark real wages are protected (see e.g. L. Bini Smaghi, ECB Board, Corriere della Sera, June 7). Now, consider the case that the central bank promises HICP = 2% over few quarters, and that therefore WI = 2%. As a result, with NDCI = 16% real costs would be rising at a pace given by 0.2´ (NDCI – HICP) = 2.8%, with output and employment being cut accordingly. The bitter discovery of wokers would be that the 2% pact with the central bank does protect real wages but at the cost of less output and employment (table 3, line 1)
Wage moderation may have another meaning concerning real wages. Suppose we wish to gauge how WI ought to be for real costs, and hence output and employment, to be sheltered against the real appreciation of NDCs. With HICP = 2%, and NDCI = 16%, the short answer is WI = 1.25´ HICP – 0.25´ NDCI = – 1.5%, that is nominal wage deflation, which amounts to a real wage cut of 3.5%. Hence the other side of the coin is that employment protection would come at the expense of the real wage.
All in all, workers face a trade-off between employment and real wage underneath the stagflation veil. Note that this result is fully consistent with the “structuralist” view of stagflation recalled above, that is, when stagflation conveys a permanent adverse change in the relative cost structure or a real supply shock (see e.g. Erceg, Henderson and Levin, 2000). The corollary is that the central bank can do nothing to mitigate this unfortunate situation (apart from letting the exchange-rate appreciate to reduce the domestic impact of imported inflation). It may well seek to achive its inflation target for general welfare purposes, but it ought to abstain from marketing this policy as being neutral.
Is the central bank’s inflation benchmark credible?
The next question is: how can the central bank deliver its promise when it cannot control the non-domestic component of inflation? The answer has been given by the ECB Board member L. Bini Smaghi (Corriere della Sera, June 7): domestic price inflation should fall below 2%. By how much can be gauged from table 2. According to our formula, given that 0.2xNDPI = 1.3%, HICP = 2% requires 0.8xDPI = 0.8% or DPI = 1%. Since the observed trend is DPI = 2.1%/0.8 = 2.6%, the deflationary pressure that the ECB should exert on the euro-economies in the coming months is in the order of 2/3 of current DPI (- 1.6/2.6). Is this a credible promise (threat)? In RGE-Monitor one can read the following statement by Trichet: “HICP inflation rates are expected to remain well above the level consistent with price stability for a more protracted period than previously thought” (www.rgemonitor.com/168?cluster_id=4680, July 16).
Hence, with WI = 2% and HICP at about 3.5 workers can expect a real wage loss of 1.5 per year “for a more protracted period than previously thought”. At the same time, real costs would rise by 1.3%, less than a half than the case in which WI = HICP = 2%. A “structuralist” would say that this result is nothing else than a shift along the workers’ trade-off curve (some real wage loss vis-à-vis less severe output/employment cut). Interestingly, Erceg et al. (2000), purporting this view, argued that since wages are notoriously rigid downwards, both in nominal and real terms, a central bank focussing on inflation too narrowly (i.e. delivering the promised inflation rate to workers) would interfere with the market adjustment of the real wage rate. The unpleasant side of this argument is that it seems to suggests that the central bank might let inflation go at the expense of moderate wage setters. The larger the “surprise inflation”, the bigger the real wage loss and the better the employment/output levels (table 3, line 2). Yet forcing workers to move down along their trade-off curve in this fashion is precisely the policy that has been banned from the modern central banker’s handbook. This policy not only does violate the credibility constraint that underpins the modern principles of central banking; in the present situation it also assumes a great deal of naivety on the part of workers.
The current symptoms of stagflation come from a real supply shock, that is a structural change in relative costs of oil and commodities.
This phenomenon generates a trade-off between domestic labour and capital incomes, on the one hand, and output and employment on the other.
- Monetary policy can do almost nothing to alleviate this trade-off, but central banks should a) not to conceal it, b) communicate clearly that finding the most preferred trade-off resolution is not their business but one of labour market negotiations, c) not to bend monetary policy to expedient manipulations of real wages.
- Consequently, setting a credible inflation benchmark to pin the inflationray process down remains a crucial ingredient, but credible means first of all realistic. As far as the EMU is concerned, the promise of a short-term return to the benchmark inflation rate, and hence the plea for the 2% pact with wage setters, are not credible.
- What is more credible is that a) the ECB can just seek to prevent inflation from pointing too far from 3.5%-4% (roughly the same figures as in the US), b) the rate of WI will result from labour market negotiations, but it will certainly be closer to 3.5% than to 2%, c) yet if WI will be kept even slightly below actual inflation (w > 0), the wage-price spiral will tend to peter out resulting in an intermediate combination of moderatly lower real wages and some limited loss of output and employment (table 3, line 3). Is this not by and large the FED’s policy? But here in Europe country specific conditions will matter the most, with Italy’s labour market having much less room for real wage cuts than France or Germany.