Wage and Price Setting: New Evidence from Uruguayan Firms

This paper presents new evidence on wage and price setting based on a survey of more than 300 Uruguayan firms in 2013. Most of the firms set prices considering costs and adding a profit margin; therefore, they have some degree of market power. The evidence indicates that price increases appear quite flexible in Uruguay (prices are downward rigid). Most of the firms adjust their prices without following a regular frequency which suggests that price changes in Uruguay are state-dependent, although wage changes are concentrated in January and July. Interestingly, the cost of credit is seen by companies as an irrelevant factor in explaining price increases. We also find that cost reduction is the principal strategy to a negative demand shock. Finally, the adjustment of prices to changes in wages is relatively fast.


Introduction
In recent years there has been a large increase in the empirical literature on price behavior. Following the work of Calvo (1983), Taylor (1980), Fuerer and Moore (1995), among others, understanding the microstructure of price setting allows us to better understand the way to fight price rigidity and conduct a more efficient monetary and macroeconomic policy. As new and detailed data sets become available, we observe a number of important studies on the microeconomic fundamentals of price setting by firmsmainly retailers -and their impact on inflation. This analysis allows a better understanding of the behavior, dispersion, and volatility of prices.
However, there are few studies that analyze price setting from surveys that directly ask companies regarding price formation, and most of the literature is concentrated in developed countries. In Uruguay in particular, despite recent progress for the retail sector (Borraz and Zipitría (2012)) and in wage formation (Fernández et. al. (2008)), there is no direct evidence on company's price formation strategies. In this study we use a novel data set of a survey of 307 large Uruguayan firms on price setting.
The purpose of this study is to present stylized facts about price setting in Uruguay based on a survey of firms. This new evidence must be viewed as a complement to the new literature on the topic in Latin America. However, the principal objective of this study is to raise a number of unknown questions about price formation for future research. Therefore, the objective is to generate questions as well as answers that would be useful for monetary policy design and to set the future agenda on the microeconomics of price setting. This study can be viewed as a first step to generate a new approach to analyze monetary policy options in a small open economy.
Our findings are as follows: i) prices in Uruguay seem to be more rigid than in previous studies, ii) the frequency of price change is state dependent, iii) the response of prices to wages is fast, iv) firms do not have a clear view on how to respond to unanticipated demand shocks; (more research is needed to understand better the response of firms to unanticipated demand shocks), v) firms seem to pay more attention to wages than their weight on the cost structure would justify, a puzzling behavior that might be related to the way wage negotiations are conducted, vi) there is a high degree of inertia in the manufacturing industry sector .
The rest of the paper proceeds as follows: Section 2 presents a brief review of related literature, section 3 makes a brief description of the data, section 4 presents the basic results of the survey and section 5 concludes.

Related Literature
As mentioned above, the price setting literature based on company surveys is scarce and most of it is concentrated on developed countries. Using surveys to analyze price-setting behavior of firms was initiated by Blinder (1991,1994) in the US. In his pioneer work Blinder interviewed two hundred firms directly regarding their price setting behavior to analyze the validity of theories of sticky prices. He found that prices are sticky with an average duration of three months following a shocks. The only theory that was supported by firms' responses was coordination failure.
In the case of Germany, Stahl (2005) finds that most of the manufacturing firms have market power to set producer prices. Additionally, indexation is minor. Babecký, Dybczak and Galuščák (2008) finds that in the Czech Republic firms' prices are less rigid than wages with a weak pass through of wages to prices. They also find that in response to an unanticipated demand shock, firms reduce temporary employment and non-labor costs.
For fifteen European countries, Druant et. al. (2009) find a close relationship between wages and prices and between wages and the frequency of price changes.
In the case of Canada, Amirault, Kwan and Wilkinson (2006), firms show a wide variation in the frequency with which they adjust prices. Almost 33% of Canadian firms declare price adjustments once a year or less while a similar portion adjust prices more than twelve times per year. Similar studies for Swedish firms (Apel, Firberg and Hallsten (2001)) and Portuguese firms (Martins (2006) (2006)) it was found that more than 30 % of total price changes are price decreases. Another important finding is that the degree of price stickiness seems to be higher in the service sector than in manufacturing.
In Latin America there are two studies based on firms' responses. The first one is Irregui et. al. (2011) analyze the link between wages and prices in Colombia using a survey of firms. They find that firms adjust their wages principally in the first quarter. They also find a stronger pass through of wages to prices in labor intense sectors or sectors with high labor productivity. The empirical evidence also suggests time-dependent price changes are less common than time-dependent wage adjustments in Colombia.
Based on a survey of 7,002 Brazilian firms, Da Silva, Petrassi and Santos (2016) find that prices are sticky, the cost of reviewing prices, statedependent price setting and markup pricing.

Data
Our study is based on a survey that was conducted by the National in Uruguay with 5 or more employees. As a result of the merger of the two samples we end up with a sample of 307 firms. Figure 1 shows the distribution of firms by sector of activity.

Empirical Results
This section presents the main results of the analysis of the survey on price setting practices in Uruguayan firms. We present the data without the weights both because of a very uneven non-response rate among sectors and because their use does not change the results.

Price setting behavior
4.1.1 Market microstructure and price setting.
We asked firms what their strategy was for setting prices. Figure (2005)). Also, the only economic sector that does not set their price with a profit margin is the transport sector, which has their prices regulated by the Government (Figure 2b). The price setting based on cost and a mark-up is highest in the trade and other business sectors. As expected, manufacturing is the sector with the highest exposure to international competition.
In Figure 2c we analyze price setting within manufacturing. Cost and mark up price setting is predominant in heavy industry. This sector also has the lowest exposure to international competition. This result reflects the high participation of heavily protected industries. Other manufacturing sub sectors tend to find the international price more important as a reference for price setting, probably as a result of lower redundant protection. If we consider domestic and international competition, the sub sectors that show a higher exposure to it are food and wood, both basic export commodities in the case of Uruguay (60% and 67% of firms respond to competition). Since the main export of goods from Uruguay are those produced by the food sector, the lower response to competition the sector shows compared to wood, which is puzzling in principle and deserves further research, might be related to market segmentation in some strategically important food components. 1 Overall, the high percentage of firms that follow non competitive practices might be related to the trade protection structure.

Frequency of price adjustment
Analyzing the frequency of price changes, Figure 3 indicates that 40% of firms do not have a regular frequency and in 30% the adjustments are semiannual. This result, based on a survey of producer and consumer firms, suggests that prices are more rigid than in the findings of Borraz and Zipitría (2012). They find that the median duration of prices in food, beverages and personal products in the retail sector is approximately two and a half months. The large proportion of firms claiming not to have a regular frequency of price adjustment might indicate that price adjustment opportunities arise in a random way, as in Calvo (1983). This result comes in stark contrast to the relatively large importance given to wages in the price formation process, particularly when wages, since Uruguay returned Our results suggest more frequent price adjustment than in Blinder's survey (1994) in the US. However, at the time of the survey, inflation was 8.9% annually in Uruguay, but at the time of Blinder's paper was 3% annually in the US. Because the pressure to change prices is lower in a low inflation environment, the difference in the inflation levels between US and Uruguay can explain the discrepancy in the results. Table 1 analyzes the correlation between the frequency of price change and employment (in thousands) and sector of activity. The dependent variable is a discrete one with values from 1 (weekly prices changes) to 7 (no regular pattern). Surprisingly, the results show a negative correlation between price rigidity and employment. However, this correlation becomes 1 The Uruguayan government has special regimes for some food commodities that have an important share of the consumption basket of the population.
positive at the employment level of four thousand. Additionally, once we control for employment, we do not find a significant correlation between the frequency of price changes and the sector of activity. This evidence seems to suggest that the frequency of price change in Uruguay is state dependent. This is consistent with Borraz and Zipitría (2012) who find that the empirical evidence seems to point to state-dependent models as the main explanation for the inflation phenomenon in Uruguay Firms stated conduct in terms of seasonality of price adjustment is barely consistent with the marked seasonality of inflation observed in the data.
In this sample most firms do not declare to have a clear pattern of seasonal price adjustment.
Only 29% of firms (see Figure 4), mostly transport and real estate sectors, declare that they change their price in a particularly month. The other firms do not concentrate their price changes at a specific time of the year. The percentage of firms that do not change prices in a particular month ranges from 50% for transport and communications and others business to 97% for the trade sector. For the manufacturing industry the response is 82%. Not surprisingly, the most important months for price adjustment are January and July, which coincide with the dates of adjustments of most of the sectors in the Wages Councils ( Figure 5). Another interesting finding obtained from the survey is that finance costs do not affect prices in any economic sector. This might reflect the fact that Uruguayan firms show relatively low levels of banking credit (see BCU (2013)). Other factors that do not have an important influence in the price of any economic sector are the price of competitors, the price of the dollar and demand. Considering that firms set their price with a mark-up we believe it is reasonable that they do not take into account the price of their competitors.

Factors Affecting Pricing
Inflation is a factor that is considered very important for transport and real estate firms at the time of changing prices. On the other hand, manufacturing industry and the trade sector do not consider it significant.
A striking fact is the high importance given by employers to wages to determine price increases. The high importance of wages in price formation contrasts with the relatively low participation of wages in the cost structure. As can be seen in Table 2 the weight of wages on total cost averages less than 20 % while raw materials are close to 60%. One can think that this is a strategic behavior by firms because of the existence of Wage Councils that mandate wage negotiations between the employers, employees and the government. Therefore, it is possible that firms overweight the importance of wages to increase prices. In order to check this, we compare the firm's response with the true structure of costs from INE for the manufacturing sector. If that is the case, this large importance of wages would be an indication of the role of aggregate demand in price formation. As wages are adjusted at the same time, firms know that the dates of wage increases (January and July) are points in which aggregate demand would jump in response to the increase in household income. The contingency analysis indicates that the correlation between the perceived importance of wages and the share of wages in total cost is positive and significant but it is far away from perfect. 2 Additionally, because price setting is costly and price are changed frequently when inflation is 8.9% (inflation in Uruguay when the survey was taken), it probably makes sense to adjust prices based on less than full information. Therefore, wage may play an important role because they may provide information about costs, demand inflation and the state of the economy.

Forward and Backward Looking Behavior
One very important question regarding price formation is the relative role Since manufacturing is the sector in which we have more data, we can see the time orientation of the sector in terms of each fundamental. Figure 9b shows that the marginal time orientation of manufacturing firms is backward looking as well.

Prices and Wages
We study the speed of price adjustment after an increase in salaries in The fact that we previously find that firms do not have a regular schedule to setting their prices is not contradictory with this fast pass trough of wages to prices. Maybe is a matter of how the firms' managers interpreted the survey. The firms do not have a calendar date for changing prices, but, they can anticipate that the prices will need to be changed after wage negotiations. Therefore, prices are not really reset that often, but firms do respond quickly because they can anticipate wage changes.

Firms reaction to an unexpected demand shock
The majority of firms, when faced with a decrease in demand, tend to reduce their costs ( Figure 11). Another reaction is to disseminate the mark-up they generate, although when they have a decrease in demand they do not decrease their prices or their production. Alike Blinder (1994), we find that firms response is not symmetric to demand or supply shocks or to negative or positive shocks. Table 3 presents the contingency analysis of the correlation between the different strategies of firms to a demand shock. Suspiciously there is a highly significant positive correlation between two opposite strategies like price increases and prices reductions. This can be explained by the fact that both strategies are not relevant at all for the firms under a negative demand shock. This result suggests a certain amount of price rigidity.
We find that the optimal response of the firm to a negative demand shock is not just one but a mix of strategies such as reduction of costs and margins and to some extent production.

Conclusions
This study provides new insights into price setting in a small economy like Uruguay based on a survey of firms. The results indicate that prices are more rigid than previously thought and indicate a relative low degree of competition in the markets.
Firms stated conduct in terms of seasonality of price adjustment is barely consistent with the marked seasonality of inflation observed in the data.
In this sample most firms do not declare to have a clear pattern of seasonal price adjustment.
Another puzzle we find in the data is the large importance they give to wages in price adjustment, which stands in stark contrast with the relatively low participation of wages in costs. This could be an indication that firms anticipate aggregate demand pressures through the behavior of wages. Also, wages may play a larger role because they may provide information about the state of the economy.
There is also a blunt contrast between the high importance given to raw materials and the dollar in price formation. This could then be the result of lower indexation due to inflation stabilization and a floating exchange rate. This stylized fact is consistent with the fall in persistence in inflation documented in Ganón (2012) among others.
Another encouraging find for monetary policy management is that firms seem to give the same importance to the past and expected behavior of fundamentals in price formation. The higher the role of expected factors, the more active the expectations channel of monetary policy is.
Finally, the firms' main response to a negative demand shock seems to be to lower costs (raise productivity).       (25)