The Highs and The Lows- Swedish Bank Failures through Inflation and Deflation (1914-26)

by Seán Kenny, Anders Ögren, and Liang Zhao

Paper available here

In the current context, it is difficult to imagine any policy discussion that does not involve interest rate movements and rising prices. The dual shock of supply chain disruptions induced by Covid-19 and the ongoing conflict in Ukraine have led to inflation rates that have not been a prevalent feature of most OECD economies since the 1980s. It is impossible to accurately forecast the duration of the current war, or establish how high central bank rates will go, but the fall in price levels that policy makers hope to achieve, and the real interest rate rises that may temporarily transpire, will not approach the extremities witnessed in the aftermath of the Great War (1914-1918).

Even in a neutral country like Sweden, price levels had more than doubled since the outbreak of hostilities, while in the post war period, it’s return to the gold standard at pre-war parity necessitated substantial price falls. As Table 1 conveys, a combination of interest rates increases and falling prices resulted in a real interest rate swing of 65 percentage points between 1918 (-40%) and 1921 (25%). While inflation ran at almost 50 per cent in 1918, the early 1920s were characterized by double-digit deflation.

Table 1: Real Interest Rates in Sweden, 1913-26

Sources: Inflation rate is from Edvinsson and Söderberg (2012); Discount rate is calculated as the annual average for the Riksbank’s discounting of bills of exchange up to three months and is from Sveriges Riksbank (1931) pp. 137 -138 and Sammandrag af Bankernas Uppgifter 1924 – 1926. Rounding may affect decimals.

These reversals, combined with weaker post war international trading conditions, contributed to produce the most severe banking crisis that Sweden had experienced, resulting in the failure of 19 out of a total of 49 existing banks. In new work (forthcoming), we study the population of Swedish commercial banks in the post war era and attempt to identify the balance sheet characteristics that enabled some banks to survive and consigned others to failure through such extreme price reversals. Unlike the traditional accounts that emphasize 1921 as the crisis year (Schön, 2010; Hagberg and Jonung, 2009), our work finds that the banking crisis materialized two years earlier and even lingered until the mid-1920s, when a second wave of banks failed after the return to the gold standard in 1924 and the currencies of other countries strengthened against the Swedish krona, as they too returned to the prewar standard (Figure 1).

Figure 1: Classification of Swedish Commercial Bank Exits (1914-27)

Source: See Paper. Note: Exits include mergers, takeovers, closures and liquidations.

The merger wave that dominated the war years is documented in our appendix, where evidence is presented on the nature of 75 bank exits between 1914 and 1926. We conclude that over 30 per cent of mergers were motivated by financial distress. We also find that aggressive takeover strategies during the boom period were weakly associated with bank demise during the crisis.

The wartime inflationary boom came to an end in 1919. During periods of inflation, the value of savings is generally eroded. Conversely, borrowers with fixed term rates see the real value of their debt eroded. In periods of deflation, the situation is reversed.  Inspired by the events of the 1930s in the USA, Irving Fisher’s (1933) ‘Debt Deflation Theory of Great Depressions’ attempted to explain the mechanism of how deflation interacts with outstanding debt. Essentially, in a world of falling prices, the borrower requires a greater volume of production or sales in the current period to repay the same amount of debt as was required in the previous period- the real value of debts rise.  As each agent attempts to sell assets or produce more (into a shrinking market) to repay or “liquidate” debt, prices fall even further, aggravating the problem. Furthermore, expectations of further deflation tend to postpone consumption, putting further downward pressure on prices. For Fisher (1933), “debts due at once are more embarrassing than debts due years hence”. Essentially, longer dated debt was preferable, as prices and the economy are afforded more time to stabilize. In contrast, shorter-term debt in a deflationary spiral would prove especially problematic, in the absence of liquidity. Schön (2010, p. 252) summarised the Swedish “deflationary crisis”- “markets disappeared, the real value of […] debts rose and additional credit became unavailable.”

Using logit regression (survival analysis) techniques on financial and corporate data of the Swedish banking population (1919-26), our study largely supports these claims. Banks are a particularly useful unit through which to view the theory, as not only do their own liability structures form a source of vulnerability in the financial system, but the “distribution in time of sums coming due” on their assets (customer loans) offers insight on whether the same mechanism affected their customers and society at large. We find that banks that financed their expansion though short-term borrowing and banks that provided higher shares of short-term credit to customers were indeed more prone to failure- debt maturity mattered most and debt deflation indeed pervaded the economy at large.

Similar to other studies of the period (Colvin et al, 2015), we found that banks that had expanded their leverage relative to their peer group during the inflationary period were more likely to fail in the downturn in their “stampede for liquidity” as they sold assets to service their maturing debts. Figure 2 suggests that such was the case during the 1920s crisis in Sweden. Those banks that grew their lending at the expense of liquidity during the boom undertook the largest deleveraging during the downturn.

Figure 2: ‘The Stampede to Liquidate’- Growth in lending/assets 1913-18 v 1919-1923

Source: Uppgifter om Bankerna. Note: Authors’ Calculations. Sample includes only those banks that existed during entire interval.

Looking at lending categories, we observe that during the boom period “risk shifting” (Turner, 2014) or “marginal lending” (Claessen et al 2014)- namely, increasing shares of lending against weaker securities (e.g. “loans against signature only”)- was a consistent determinant of distress.  Banks that had reduced their share of lending against property also suffered, as such longer-term loans were better able to weather the price gyrations of the early 1920s.

Finally, our study reviews an overlooked aspect of the 1920s crisis: foreign exchange reversals were a primary culprit in explaining Swedish bank failures, particularly in the later phase. Many Swedish banks had financed their wartime expansion through foreign loans as the krona strengthened. As Figure 3 illustrates, the situation reversed dramatically in the post bellum period and the value of foreign debt (in krona) rose sharply, resulting in a number of bank failures. The process was repeated in 1924 as the Swedish krona weakened against the Danish and Norwegian currencies, exposing a final wave of debtor banks to distress.

Figure 3: Exchange Rates in SEK, 1913-27

Source: Bohlin (2010). Note: GBP and USD on right axis.